Abstract
Purpose: :
To present an ontology used in conjunction with a database for integration of clinical and imaging data, that can allow a faster approach to data entry by physicians and extended collaboration between different sites through VPN connections.
Methods: :
The database, intended for clinical use, is developed on a modified three tiers model, in which the client interface was designed in Microsoft’s FoxPro and C++, object oriented languages, with a SQL 2005 Server back-end support for security and scalability. The system resides on a Windows 2003 Server Standard edition with Terminal Services. For added security and easy upgrades the terminal client connection is used for access from any individual computer. In this way we benefit from all security provided by the default encryption of terminal services. No information resides on individual workstations. The system is automatically disconnected after 15 min of inactivity. Password access is used at this time, and we established HIPPA compliance through internal procedures, in place at our institution.The database consist of one main table with demographic information and multiple children tables with clinical visits, imaging and treatment information. The design took into account flexible structure for modifications and upgrades. The screens are ordered accordingly to subject in the clinical observation charts. We opted for a free text entry system that is controlled by specific ontology developed for that specific entry, like patient history or ocular history or medication used. All these area have a limited controlled vocabulary or ontology what can serve as a check point for the consistency of data acquired. The program can store or link images from different image acquisition subsystems, and in a future development tags would be used for representing image content. Previously presented algorithms for image enhancement and segmentation can be applied through an integrated analysis interface.
Results: :
Different report generation features, for complex queries, were implemented, making easy for clinicians to have a comprehensive analysis done at a press of a button. Also the system can export all the data or specific queries in different formats including XML. The export also can be done in HIPPA compliant deidentified format.
Conclusions: :
A system with this design can be used with hand writing or voice recognition tools because of the better handling of acceptable input.
Keywords: imaging/image analysis: non-clinical • computational modeling