The treatment of certain diseases such as cancer, HIV, or other serious medical conditions relies on a regular administration of critical drugs that are necessary to keep those life-threatening diseases under control. Those drugs (e.g. Efavirenz, Imatinib, Tacrolimus, Tobramycin) have a narrow therapeutic range and a poorly predictable relationship between the dose and the blood drug concentration, which may vary greatly among individuals. Therapeutic Drug Monitoring (TDM) aims at improving patient care by monitoring drug levels in the blood to individually adjust the dosage for targeting drug concentration in the therapeutic interval. In order to ensure a better prediction of the relationship between dose and drug concentration, the ISyPeM2 project (a continuation of the Nano-Tera project: Intelligent Integrated Systems for Personalized Medicine, ISyPeM, http://www.nano-era.ch/projects/368.php) has developed a Bayesian TDM approach [GWM+12] based on studies in general or special populations. This approach requires population health data (covariates, dosages, drug concentrations) to be collected and analysed by researchers, in order to enhance the prediction models. Therefore the following question arises: how is it possible to share and aggregate medical data for research purposes while preserving the patients’ privacy?