A framework for a clinical reasoning knowledge warehouse

Abstract

In many areas of the medical domain, the decision process i.e. reasoning, involving health care professionals is distributed, cooperative and complex. This paper presents a framework for a clinical reasoning knowledge warehouse that combines theories and models from artificial intelligence, knowledge management systems and business intelligence to make context based, patient case specific analysis and knowledge management. The knowledge base integrates three sources of information that supports clinical reasoning: general information, guidelines and health records. New generalized knowledge is stored and made accessible when relevant to the reasoning context and the specific patient case. Furthermore, the information structure supports the creation of new generalized knowledge using data mining tools. The patient case is divided into an observation level and an opinion level. At the opinion level, reasoning participants can express their argument based opinions about a patient case, thereby enhancing the knowledge about the state of and plans for the patient. An opinion language that supports expressing a possible imprecise/uncertain opinion based on imprecise/uncertain/incomplete arguments is introduced. Finally, case based reasoning is used for retrieving similar patient cases and adapt the opinions in the current case.