Evidence based medical research requires high quality data from health care provisioning. However, necessary data sets are frequently not easily available for medical studies. Data sets might be fragmented over databases from numerous health care providers. They are heterogeneous and are stored in various different formats. They are not collected using standardized protocols and data quality specifications. Their use might be restricted by privacy requirements. Informed consents for their use might be inconsistent. Access procedures might be complex and bureaucratic. We analyze these problems and discuss different approaches for setting up infrastructures to facilitate the use of health care data for medical research: in particular, privacy preserving repositories for identifying relevant datasets for medical studies, wholesome data quality management systems, and strategies for enabling secondary use of both health care data and data derived from medical studies.