Research On The Veteran TCM Physicians' Scheme Of Syndrome Differentiation And Treatment Of Chronic Gastritis Based On Knowledge Graph

Objective:1)Based on multi-source and heterogeneous data sources,the knowledge graph technology is used to construct the inheritance graph of veteran TCM physicians'experience,which provides a visual tool for analyzing and displaying the physicians's clinical thinking,diagnosis and treatment rules,inheritance characteristics,and lays the foundation of knowledge base for clinical decision support system of TCM.2)Based on the real world clinical data,a scheme of syndrome differentiation and treatment based on the evaluation of curative effect is constructed,and the law of syndrome differentiation and treatment of chronic gastritis by two doctors with teacher-student relationship is compared,so as to provide methodological reference for the systematic process of making individualized diagnosis and treatment scheme of TCM.Method:1)Research on the construction of graph database of veteran TCM physicians experience:Taking the clinical experience of Dr.Yao Naili as the research object,26 pieces of unstructured text data are systematically retrieved,and build the ontology based on the Categories of traditional Chinese medicine(TCM)clinical terminological systems(ISO 19465-2017)and the Health informatics-Semantic network framework of traditional Chinese medicine language system(ISO/TS 17938-2014).The data entity is extracted according to the formed knowledge ontology,the data layer of knowledge graph is formed through knowledge fusion,knowledge representation and other links,and the data storage and application display of knowledge graph are realized through Neo4j graph database.2)Study on the law of syndrome differentiation and treatment of chronic gastritis of veteran TCM physicians and its inheritance characteristics:Taking Dr.Yao's treatment of chronic gastritis as an example,using retrospective hospital information system outpatient medical record data,on the basis of evaluation of clinical efficacy,through descriptive statistics,data mining and other ways,to clarify its effective scheme and appropriate population characteristics,and compared with his student Dr.Zhang Runshun,to reveal their own characteristics of syndrome differentiation and treatment,so as to lay a foundation for the next step to form a personalized diagnosis and treatment scheme.3)Research on the diagnosis and treatment scheme of veteran TCM physicians based on expert interview:Based on the real world diagnosis and treatment data mining results,this paper discusses the diagnosis and treatment ideas of the two doctors by qualitative interview research method,aiming to form the individualized clinical diagnosis and treatment scheme of chronic gastritis,and realize the function of expert diagnosis and treatment scheme recommendation by using knowledge graph technology.Results:1)The construction of knowledge graph of Dr.Yao's experience in diagnosis and treatment:Based on the semantic types and relations determined by Dr.Yao's text data,the graph of Neo4j mode layer is constructed,including 19 nodes;18 labels,27 relations and 17 types of relations.The data layer of knowledge graph is formed by knowledge extraction and knowledge fusion.There are 524 nodes under 18 lables,1097 relationships under 17 relationship types and 622 property keys under 10 properties.Finally,the knowledge graph of Dr.Yao's diagnosis and treatment experience realized the function of visual display and semantic search.2)Study on the rule of syndrome differentiation and treatment of chronic gastritis and inheritance characteristics of Dr.Yao:1176 cases of chronic gastritis were included after curative effect evaluation,including 356 cases of Dr.Yao,2383 visits;820 cases of Dr.Zhang,4632 visits.Through data mining methods such as community detection and complex network analysis,the core prescriptions and symptom groups of two doctors for chronic gastritis were formed.The core prescription of Dr.Yao is based on the self-made Jianpi Tongluo Jiedu formula,and the main complaint is stomach pain;the prescription of Dr.Zhang is based on Sini Powder,Banxia Xiexin Decoction,Xuanfu Daizhe Decoction,and the main complaint is stomach distention and belching.3)The formation of TCM diagnosis and treatment scheme and the optimization of knowledge graph:Through qualitative interview research,Dr.Yao's interview text totally extracts 5 codes,which are syndrome differentiation and treatment,disease differentiation and treatment,combination of disease and syndrome,basic pathogenesis and prescription;Dr.Zhang's interview text totally extracts 7 codes,which are disease and syndrome combination,main disease prescription,diagnosis key point,patient type,combined symptom,basic thinking and prescription.There are two preset themes,diagnosis and treatment.Finally,the individualized diagnosis and treatment scheme for chronic gastritis of two doctors was formed,and was presented in the form of knowledge graph,realizing the function of recommendation of expert diagnosis and treatment scheme,and achieving the purpose of assisting clinical decision-making.Conclusion:1)Knowledge graph technology is one of the effective methods to study the inheritance of veteran TCM physicians.The knowledge graph realizes the functions of visual display,semantic search,treatment recommendation and so on,which lays a methodological foundation for knowledge representation and knowledge engineering in this research field in the era of big data.2)Based on the real world clinical data of efficacy evaluation,combined with quantitative and qualitative research methods,the research path of establishing the diagnosis and treatment scheme of veteran TCM physicians is basically feasible.3)This study has formed 4 superior schemes of syndrome differentiation and treatment recognized by doctors themselves,which provide a practical,stable and reliable diagnosis and treatment scheme for the future clinical treatment of chronic gastritis in TCM.