Why you might have multiple configurations

There are a variety of public and company-specific pharmacovigilance databases that share a common conceptual organization (case reports that contain information on demographics, drugs, and events) but differ in details of field and table naming, presence or absence of specific attributes, and use of specific medical coding dictionaries. These configurations can be set up by expert users who are familiar both with the target database schema and the pharmacovigilance application; medical end-users work with the application-specific variables so defined (end-users do not need to be aware of the target database schema). Multiple configurations can support user access to several different databases (e.g., to do a data mining run—first, on the public data and, then, on in-house data), and also to several different versions of the same database (e.g., different chronological snapshots or versions that include or exclude so-called “concomitant” medications).

To insulate both Oracle Empirica Signal and users from superficial variation in the detailed formatting of the target safety database, Oracle Empirica Signal supports multiple database configurations, each of which defines a mapping between user-visible “variables” and specific database tables and columns, including specification of plausible roles for the variables (e.g., a drug name, an event name, an attribute suitable for use in stratification or subsetting, etc.).

When Oracle Empirica Signal is installed, a master configuration table that contains all configurations for all source databases is created in another Oracle account. This master configuration table is populated automatically as configurations are added, copied, or imported. When a configuration is created, a configuration table is created automatically in the same Oracle account as the source data. Multiple configurations (and configuration tables) may exist for an Oracle account.