Towards Private Medical Data Donations by Using Privacy Preserving Technologies

Through the growing amount of personal health data collected by the individual itself digital data donations become more and more attractive. Wearables like AppleWatch or Fitbit trackers make tracking of heart rate, daily step counts and other lifestyle data easier than ever. While this data is collected on the dedicated device, it can help research in many promising ways. Even if the potential benefit of this data is very clear, there are open questions regarding privacy. Traditional privatization measures like anonymization and pseudonymization can only provide limited privacy guarantees especially with the growing amount of personalized data. To mitigate those risks privacy enhancing technologies like differential privacy can be used. While the theoretical foundation of such technologies is strong, only limited data is available about their practical use in large scale applications and the trade-off between privacy and utility. In this paper we will present a data donation scenario that is inspired by a real-world use case using lifestyle data for its analyses. We will apply the local differential privacy technology "RAPPOR" to improve the privacy protection for the data donors and evaluate the impact of this technique to the data utility.