2014 Poster Sessions : Efficiently Discovering Privacy-Leaking Association Rules in Large Medical Discharge Databases

Student Name : Peifung Eric Lam
Advisor : John Mitchell
Research Areas: Theory
Abstract:
Patient privacy protection can encourage patients with embarrassing but serious sensitive medical conditions such as STDs, substance abuse or mental health disorders to seek necessary medical help without fearing stigma. We present an initial study on what common patterns of medical codes may suggest the presence of a sensitive condition to someone outside a patient’s care team, even if the primary codes have been masked from the original record for privacy protection. We develop an algorithm to find such associations efficiently in large medical databases.

Our study suggests that some sensitive chronic disease conditions may co-occur with patterns of related medical codes that can be more identifiable to parties with access to medical domain or statistical knowledge. We evaluate the risk of such inferences and discuss techniques to defend against them.

Joint work with Ellick Chan and John C Mitchell

Bio:
Peifung E. Lam is a Ph.D. candidate advised by Prof. John C. Mitchell at the Computer Science Department, Stanford University. His research interests include formal models on the security and privacy issues in the spaces of web security, privacy regulation compliance, and healthcare privacy.