What Information Does Your EHR Contain? Automatic Generation of a Clinical Metadata Warehouse (CMDW) to Support Identification and Data Access Within Distributed Clinical Research Networks

Data dictionaries provide structural meta-information about data definitions in health information technology (HIT) systems. In this regard reusing healthcare data for secondary purposes offers several advantages (e.g. reduced documentation times or increased data quality). Prerequisites for data reuse are its quality, availability and identical meaning of data. In diverse projects, research data warehouses serve as core components between heterogeneous clinical databases and various research applications. Given the complexity (high number of data elements) and dynamics (regular updates) of electronic health record (EHR) data structures, we propose a clinical metadata warehouse (CMDW) based on a metadata registry standard. Metadata of two large hospitals were automatically inserted into two CMDWs containing 16,230 forms and 310,519 data elements. Automatic updates of metadata are possible as well as semantic annotations. A CMDW allows metadata discovery, data quality assessment and similarity analyses. Common data models for distributed research networks can be established based on similarity analyses.