Research Of Clinical Knowledge Expression And Parse Based On Multi-agent

With the development of society, medical information has become a trend thatmust follow. Restricted by many conditions from different aspects, our country started alittle late on the way to medical information and is facing many problems andchallenges now. One of the important problems is that the current clinical knowledgeand practical experience has accumulated enough, but cannot be used efficiently,because they are scattered at different places. If we can organize them togetherefficiently and provide them for medical workers in a timely manner, thus could helpclinicians make more scientific and much better decisions. Then the current quality ofmedical services would be further guaranteed and the relationship between doctors andpatients would be more harmonious at the same time.Clinical guidelines summarize the newest clinical knowledge and experience andhave higher practical value. The research content of this paper is that transform thewidely accepted clinical guidelines into knowledge documents and then put them intothe use of clinical decision support systems through the knowledge interpretation engineto assist doctors make better decisions. Based on the research content, this paper didwork in the following three aspects:(1)Define the forms of the expression of clinical guidelines and then transform theguidelines into knowledge document in the XML format that will be stored in theknowledge repository through the clinical knowledge acquisition tool, which hasalready been developed by other team members.(2)Design and implement the knowledge interpretation engine. With the parse ofthe knowledge document by using this engine, it will generate web-based assisteddiagnosis pages dynamically. The pages will prompt doctors what key symptoms haveto ask or what tests have to be taken during the diagnosis process and will finally givepersonalized diagnostic conclusions or decision advice based on the match situationbetween clinical knowledge and the patient’s symptoms.(3)Extend the work of clinical knowledge expression and parse to a distributedenvironment to achieve cross-medical specialty medical cooperation pattern byintroducing the rule-driven based multi-agent technology. In this pattern, many agentsparticipate in the disease consultation and complete the diagnosis and treatment ofdisease through mutual cooperation. Every agent will be assigned tasks according to itsrole. Agent will complete the tasks by taking advantage of its own knowledge that is inthe format of rules. Agents are ordered to complete their tasks to complete thecooperation of disease diagnosis and treatment.This paper has verified the legitimacy and effectiveness of clinical knowledgeexpression and resolution by using the diagnosis of urinary tract infection in children asa case and has proved feasibility and scalability of rule-driven based multi-agenttechnology that used in the distributed knowledge expression and application by using the breast cancer referral as a case.