Air Force Consolidated Cancer Epidemiology Repository (ACCER)

Cancer epidemiology research on military personnel is hampered by the fact that the necessary data is stored in separate databases. Researchers must manually link information from tumor registry, manpower, and personnel databases. Our solution proposes to fuse such data in an automated fashion. We will explore creation of a customized data cache containing a normalized collection of relevant data points from the data sources. This cache must be based on a metadata standard, designed to ensure semantic and syntactic comparability across the data. Research will be necessary to determine the best methods of making the data cache structure independent of the specific database definitions. We will design a data extractor and data normalizers that are generic in function yet work with each data source. Researchers will be able to select cohorts and download data through a straightforward web-based user interface. Once the framework is developed it may be extended to any number of additional data sources. The Scenario-based Engineering Process, which is ideal for projects with non-obvious requirements, will drive our approach. Frequent shareholder input is key to the process and ensures that all relevant issues are uncovered and addressed.