Huge amounts of Personal Health Information (PHI) are increasingly being available, used, and published for secondary purposes such as research, public health, and marketing. Along with such increasing secondary usage of PHI, privacy concerns among public have been raised due to the potential privacy breaches which may arise from improper release and usage of person-specific health data. To address these concerns and to prevent privacy violations one promising approach is to anonymize personal health data. Data anonymization generally resides in the context of privacy preserving data publishing (PPDP), but there are some specific requirements for de-identifying health data which makes it more challenging and most of the available approaches in PPDP will not be practical for use with health data. Moreover, health data, particularly the clinical data, are often longitudinal by nature and, to the best of our knowledge, no methods have been proposed so far for de-identification of longitudinal clinical data. In this research, we are developing methods for publishing longitudinal clinical health data so that the disclosed data remain practically useful in secondary use contexts while the privacy of individuals is preserved. We are particularly interested in privacy models which thwart attribute disclosure attacks.

* This work is part of a research project at the Electronic Health Information Lab, Children's Hospital of Eastern Ontario Research Institute, Ottawa, Ontario, Canada