Research Of Multi-agent Support Framework In Distributed Clinical Decision-making Support System

Medical errors have become a common concern of social problems, notably complex diseases. One important reason is the lack of clinical experience. Clinical guidelines are the solution to this problem; however, they are always kept being updated. Let the system adapt to the changing clinical guidelines is a challenging idea. In addition, if the formation of a care team to treat this patient who was afflicted with cancer, chronic disease and other complex diseases, can greatly increase the probability of disease cure. In this paper we propose Multi-Agent Support Framework in distributed Clinical Decision-making Support System can achieve these goals. In the MAS, clinical guidelines are mapped into rules, which are stored in the rule repository and could be processed by the Execution Engine, to guide agents’ behaviors. Rules in the system are configurable and Execution Engine can always obtain the latest data at runtime without interrupting system running, so it implements the system’s adaptability. Each Agent has its own tasks and functions and they communicate with each other through Execution Engine to implements data sharing, expands the scope of knowledge sources and improves the accuracy of decision-making greatly.The MAS has the following advantages: Firstly, a change of system requirements do not affect the executing of MAS and so improve the adaptability of the system; Secondly, every agent can complete their own tasks and functions through performing the same set of Execution Engine, reducing the amount of code, improving the efficiency of the system; Thirdly, It is based on the distributed environment and can communicate with each other by sending the message to talk about the disease and finished goals. So it is a more realistic simulation of the human society in different physical environments to diagnose disease. Combined with the content of the paper to study, I did work from the following three aspects to implement:(1) In the MAS, clinical guidelines is described by XML format, we designed a multi-agent Execution Engine to parse this type of XML file and it can abtain these updated content at runtime timely without interrupting the operation of the system.(2) In this paper the conclusions deduced by the system shows in a user friendly interface,the doctors can choose and pick the most valuable advice for their situations or make their own decision according to the actual situation. It reduced the probability of clinical decision-making mistakes due to lack of doctor’s own clinical experience.(3) The paper presents the idea of Multi-Agent Systems, which can divide a health goal into many small parts. To implement the diagnosis of complex disease by communicating and discussing among the agents in the distributed environment. It solved the difficulty of doctor’s communication and making up a team.