A Method and Architecture for Building Compliance Agents

James Davis, Michael Huhns and Ronald Bonnell

This position paper presents preliminary results on defining a method and architecture for building agents that monitor compliance to specified plans or protocols in the healthcare domain. This work is being pursued as a precursor to constructing an automated, problem-solving task-specific, knowledge acquisition environment for agent-based compliance monitoring software applications in healthcare. The status of our work is as follows. One agent system has been developed and evaluated for treatment compliance monitoring in a disease management sub-domain, while a second agent system is being developed for compliance monitoring of occupational safety guidelines in the healthcare workplace. We believe that our work in creating an architecture for compliance monitoring problems can be used as a basis for building a knowledge-level, task-specific agent development tool set, along the lines of those created for generic-task based knowledge acquisition environments. Our current emphasis is on articulating the method devised for creating these agent systems, and its underlying system architecture, that is enabling us to create a model-driven acquisition environment for building domain-independent compliance agent applications. Instead of using knowledge-level analysis techniques for articulating our problem-solving architecture, we have adopted object-oriented analysis techniques for this purpose. We have found these useful for capturing problem semantics that are precise and rigorous, yet amenable to automation through a well-understood mediating interface required of a knowledge acquisition tool kit.

This page is copyrighted by AAAI. All rights reserved. Your use of this site constitutes acceptance of all of AAAI's terms and conditions and privacy policy.