Study Of Knowledge Representation And Application Methods For Clinical Practice Guidelines

The application of clinical guidelines for clinical diagnosis and treatment can improve the safety and quality, and its importance has been widely recognized. However, amounts of clinical guidelines are not well applied in clinical practice due to their complexity for reading and understanding. Clinical decision support systems employ medical knowledge bases, help the doctors to collect and analyze patient data for physicians and provide clinical decision-making information for the doctors'reference. Representation of the clinical guidelines with clinical guideline models can make great use of them in clinical practice. The research in this field has been developed for above 20 years and produced many clinical guideline models for certain knowledge representation. However, the research of representation for uncertain knowledge has been developed less owing to its complication. Currently, there is short of a well-developed clinical practice models that can simultaneously express certain and uncertain knowledge.Therefore, this article focuses on the implementation of clinical knowledge representation methods for clinical guidelines. First, we represented uncertain knowledge in clinical guidelines with fuzzy logic and decision tree according to the successive layers in clinical guideline knowledge. It is called fuzzy logic model. Simultaneously, we choose SAGE model as the certain knowledge model in this paper.In order to ensure that they can work together, clinical events in HL7 are applied for representing the clinical decision service in the network and structured levels of both models. vMR model is employed as the standard interface for medical data service. We also established the clinical workflow representation method to ensure the physicians can call different clinical decision support services flexibly under the actual needs of clinical practice. Consequently, SAGE model and fuzzy logic model can work together in the same clinical guideline model called muti-model clinical guideline representation framework.To verify this model and put it into clinical practice, we developed two clinical cases according to the model for metabolic syndrome and Alzheimer's disease guidelines with protege, MATLAB fuzzy logic toolbox and CLIPS inference engine. Clinical decision support system were also established and 200 clinical cases are used for the clinical evaluation in the hospital. The evaluation has shown positive results in the improvement of clinical levels of diagnosis and treatment.