Seven Ways to Evaluate the Utility of Synthetic Data

Abstract

Access to individual-level health data is going to be critical for managing the COVID-19 pandemic and enabling society to return to some form of (new) normal functioning. Broader data access is already starting to happen. At the same time, there has been growing alarm by the privacy community about the extent and manner of the level of data sharing that is going on with such sensitive information. In South Korea, broad data sharing has already resulted in some patients being reidentified and experiencing judgment and ridicule,1,2 and some governments have begun to reduce the amount of information being shared about COVID-19 cases. 3-8 Data synthesis can provide a solution by enabling access to useful information while ensuring reasonable privacy protections.