Agent Based Integration Model: A Case Study Of Kra And Its Medical Service Providers

Abstract

Kenya Revenue Authority provides medical care services to all its members of staff and their dependants through accredited medical service providers. These medical service providers use their own patient management systems to manage patient records. A member of KRA staff or beneficiary can visit any of these medical facilities and get treatment, after which a hardcopy invoice is sent to the authority for settlement. Real time invoices are not available and the physical ones could even take a month to get delivered to the Authority. On the other hand, the service providers do not have a proper mechanism of validating the staff or beneficiary before treatment can commence. A prototype has been developed which validates staff members and beneficiary before the start of the treatment commences. This validation is done by the facility admission office using a web interface which the facility users must first login into, after which validation can be done against the staff ID or beneficiary’s medical card no. The patient invoices from the health centres are consolidated and relayed to KRA medical system via Multi Agents System approach, which work in the background to inform the central synchronization server of any patient treatment information taking place. Approaches that have been used previously in systems integration include building new Systems, Web services integration using XML, common data storage mechanisms using data ware houses, data replication and by the use of integration by middleware like SQL middleware. The integration approach used in this model revolves around abstraction at the data level with the aid of JADE frame work. The system agents work cooperatively with other agents in their social network that comprises heterogeneous systems comprising of syntactically and semantically differing data sources. The system delivers output by consulting and retrieving patient treatment records at a facility and where necessary updating the historical data warehouse while maintaining the ACID properties.