Using Private Data for Public Purposes Securely

Greetings from the famous room N101 at Australian National University (ANU) in Canberra, where Professor Bradley Malin of Vanderbilt University is speaking on "Towards Practical Private Data Integration and Analysis".  He is describing the problem of using medical records to detect public health problems. People's medical records need to be kept private, but access is needed to this information to detect an outbreak of disease or an adverse drug reaction. Researchers are investigating ways to protect privacy and also make use of the data, such as Homomorphic Encryption. The mathematics involved is complex, but some of the techniques are easy to understand. One is that fields which have only a few values, such as gender, need to be scrambled more than some which many values (such as address). He has implemented this in the Secure Open Enterprise Master Patient Index (SOEMPI) is software tool. Testing of the security of such techniques is traditionally carried out by trying match public information provided by government about individuals with that which is private. The most common example of public records in the USA is voter registration. However, individuals now provide much more information themselves via the Internet. Does this additional information make such attacks easier?

One of his papers is "A Practical Approach to Achieve Private Medical RecordLinkage in the Face of Public Resources".

Towards Practical Private Data Integration and Analysis

Assoc Prof Bradley Malin (Vanderbilt University, Nashville)

CECS SEMINAR SERIES

DATE: 2013-08-26
TIME: 16:00:00 - 17:00:00
LOCATION: CSIT Seminar Room, N101
CONTACT: peter.christen@anu.edu.au

ABSTRACT:
Over the past decade, it has been repeatedly demonstrated that data devoid of explicit identifiers can be linked back to the identities of the individuals from which it was derived. This has made organizations increasingly apprehensive about sharing person-specific information. Yet, with the dawn of the big data age upon us, it is imperative that data sharing proliferate to ensure that researchers can validate published research findings, combine datasets to discover novel associations, and comply with open data initiatives. In this talk, I will review recent research on privacy preserving data integration strategies that are efficient, effective, and obscure personal identities in the process. This talk will further illustrate how such integration can enable biomedical association studies while obfuscating the identities of the corresponding participants.
BIO:
Bradley Malin, Ph.D., is the Vice Chair for Research and an Associate Professor of Biomedical Informatics in the School of Medicine at Vanderbilt University. He is also an Associate Professor of Computer Science in the School of Engineering and is Affiliated Faculty in the Center for Biomedical Ethics and Society. He is the founder and current director of the Health Information Privacy Laboratory (HIPLab), conducts technologies that enable privacy in the context of real world organizational, political, and health information architectures. Dr. Malin's research has been cited by the U.S. Federal Trade Commission and featured in popular media outlets, including Nature News, Scientific American, and Wired magazine. He has received several awards of distinction from the American and International Medical Informatics Associations and, in 2009, he was honored as a recipient of the Presidential Early Career Award for Scientists and Engineers (PECASE), the highest honor bestowed by the U.S. government on outstanding scientists and engineers beginning their independent careers. Dr. Malin completed his education at Carnegie Mellon University in Pittsburgh, PA, where he received a bachelor's in biological sciences, a master's in data mining and knowledge discovery, a master's in public policy and management, and a doctorate in computer science.