Synthetic health data have the potential to mitigate privacy concerns when sharing data to support biomedical research and the development of innovative healthcare applications. In this work, the authors introduce a use case oriented benchmarking framework to evaluate data synthesis models through a set of utility and privacy metrics.