Secure Medical Data Collection via Local Differential Privacy

As the volume of medical data mining increases, so do the need to preserve patient privacy. And the exposure of medical data may degrade the level of health care service and reduce the trust of patients. Local Differential Privacy (LDP) was proposed to solve problems in the context of local privacy, by which data collectors are hardly to get exact individual information. We present a secure medical data collection framework and apply our framework on synthetic data at different scale. Finally, we evaluate the performance of our work. Our experimental results show that both privacy and validity of medical data are aligned.