To harmonize data between studies participating in the consortium, study-specific datasets and Harmonized Datasets (i.e. variables targeted for harmonization) are used and made available to investigators. Variables collected by each study and those defined in the Harmonization Dataset are described using a standardized data dictionary which covers variable definitions and characteristics (e.g. type, unit, categories, and area of information covered). Once Study Datasets and a Harmonization Dataset are defined and loaded in the system, variable-specific mapping algorithms are developed in order to derive the study datasets into the common format defined in the Harmonization Dataset. In order to increase transparency of the harmonization process, these resources are made available to the research community.
For studies using the OBiBa Opal software, the Mica application includes a distributed query engine that enables researchers to obtain real-time aggregated reports on the exact number of participants with specific characteristics or phenotypes. Mica also allows running regression analyses remotely on real study data using the DataSHIELD method (Wolfson M et al. Int J Epidemiol. 2010). More information on the OBiBa suite of software is available at www.obiba.org.