The mondriaan project: The dutch healthcare landscape as a population laboratory

Background: The gap between pre-clinical pharmaceutical research and clinical/epidemiological research hampers optimal benefit-risk assessment and monitoring of new and existing medicines. Many excellent health care databases are available in the Netherlands for pharmaco-epidemiologic and -economic research. However, in isolation these data remain scattered and have limitations with regard to sample sizes and/or detail of the registered information. Objectives: To optimize routine health care databases in the Netherlands for pharmaco-epidemiologic and economic research in order to improve benefit-risk assessment and monitoring of new and existing medicines. Methods: We build an ICT infrastructure for linkage of healthcare/research databases in the Netherlands. To protect privacy, pseudonimisation and linkage was performed by a trusted third party (TTP). A data catalogue was developed to allow queries over the integrated databases which will help in designing pharmacoepidemiologic studies (incl. sample size calculations, assessment of completeness of data). Results: We linked pharmacy records from the National Foundation of Pharmaceutical Statistics (n>12,000,000) on a patient base to several routine health care databases such as the Almere Health Care database (n=160,000), the Utrecht Health Project (n=20,000), and the AGIS claims database (n=1,200,000). Currently we are integrating several other databases in the Netherlands. Conclusions: The project will deliver a large-scale, highquality data platform for innovative pharmaceutical research and provide the tools to improve benefit-risk assessment of medicines.