CAPRA: A Cancer Patient information Retrieval Agent

1. Efficiently automate database search. retrieval. and browsing of PDO content. CAPRA mimics a human search intermediary, maintaining knowledge of the database schema and using it to tell the user what data 'destinations' are available. Once a destination is selected, it uses schema knowledge to define and display all potential search criteria, based on a dynamic calculation of all possible data paths by which the destination could be reached. (E.g., clinical trial protocols could be retrieved through sponsoring organizations, drugs used, inclusion criteria, etc.) Additional knowledge on the SQL database query language is used to search the database, and format the results for display to the user. This entire process is called database searching. Once specific data are obtained, CAPRA can identify structurally related information in the database and allow the user to retrieve and examine these related items. This process of finding related information without additional searching is called database browsing.