| BackgroundChronic gastritis is a common digestive system disease in clinical practice that is difficult to cure and has a persistent and recurrent course,greatly affecting the patient’s quality of life.Traditional Chinese medicine(TCM)treats chronic gastritis through individualized treatment based on syndrome differentiation,which can not only relieve symptoms but also adjust the internal environment of the body,enhance immunity,and improve self-regulation ability,achieving significant therapeutic effects.The TCM clinical case history is the main carrier for TCM diagnosis and treatment of diseases,reflecting the physician’s thinking on disease diagnosis and treatment.Medical practitioners can use clinical case histories to standardize clinical workflow,strengthen disease management,reduce medical errors,and improve the quality of medical services.The clinical case history contains a large amount of structured and unstructured data,and mining and analysis of this data can provide more comprehensive and accurate clinical characteristics of patients with chronic gastritis,thereby providing more effective evidence support for TCM diagnosis and treatment decision-making.Knowledge graph,as an emerging knowledge representation and reasoning technology,is widely used in knowledge management,information retrieval,and intelligent question answering.Constructing the TCM treatment of chronic gastritis diagnosis and treatment knowledge into a knowledge graph can help to deeply explore its inherent laws and knowledge structure,improving the accuracy and personalization of TCM treatment of chronic gastritis.With the advent of the big data era,data mining technology has been widely used in the medical field.Through data mining technology,the diagnostic and treatment thinking of physicians can be systematically and structurally mined to explore their diagnostic and treatment rules.Mining large amounts of medical data can reveal key factors and treatment plans that affect patient’s condition,optimize treatment plans,and improve therapeutic effects.By conducting data mining analysis on the characteristics of chronic gastritis patients,the disease-syndromesymptom characteristics,and medication rules,the diagnosis and treatment knowledge of chronic gastritis can be deeply discovered,revealing the clinical medication rules of TCM in treating chronic gastritis,providing ideas for clinical practice.The combination of knowledge graph and data mining technology can provide a more comprehensive understanding of the TCM treatment system for chronic gastritis and provide methodological reference for summarizing and promoting the diagnosis and treatment experience of chronic gastritis.ObjectiveThis study utilized real-world clinical medical records as the data source and employed knowledge graph technology to construct a TCM diagnosis and treatment knowledge graph for chronic gastritis.This provides a visualization tool for analyzing and displaying implicit diagnostic and treatment thinking,TCM differential diagnosis and treatment rules,and commonly used medicines by physicians.Additionally,this study offers insights for the development of semantic search,question-answering systems,and recommendation applications based on the TCM knowledge graph for chronic gastritis.Furthermore,by utilizing chronic gastritis patient medical records and data mining techniques,this study discovered diagnostic and treatment knowledge for chronic gastritis and compared it with the constructed knowledge graph.The goal is to further explore diagnostic and treatment knowledge for chronic gastritis and provide support for the improvement and optimization of the knowledge graph.Methods1 Research on the construction of a traditional Chinese medicine diagnosis and treatment knowledge graph for chronic gastritis based on clinical case records:This study constructed an ontology framework and a knowledge graph by utilizing the data extracted from the electronic medical record system and hospital information system of patients with chronic gastritis and referencing the diagnosis and treatment guidelines for chronic gastritis as well as commonly used textbooks.Firstly,chronic gastritis TCM medical record data is collected,and data that meets the annotation criteria is processed through entity annotation,relationship extraction,terminology standardization,and other steps through human-machine cooperation and machine learning.Secondly,with the guidance of domain experts and reference to existing literature,standards,and terminology systems,the ontology layer of chronic gastritis is designed rationally.The Stanford University’s " seven-step method" is then used to construct the chronic gastritis TCM diagnosis and treatment ontology,forming the pattern layer of the knowledge graph.Next,the organized and standardized chronic gastritis diagnosis and treatment data is matched and corresponded with the ontology framework to form the data layer of the knowledge graph.Finally,all entities and relationships are imported into the Neo4j graph database,forming the semantic retrieval query and visual display of the knowledge graph.2 Research on knowledge discovery of TCM diagnosis and treatment of chronic gastritis:The study aims to analyze the basic information,common-symptoms,signs,and use of traditional Chinese medicine of chronic gastritis patients who received treatment in the gastroenterology departments of Xiyuan Hospital of China Academy of Chinese Medical Sciences and the Guang’anmen Hospital of China Academy of Chinese Medical Sciences between January 2013 and October 2021.IBM SPSS Modeler 18.0,SPSS 26.0,and Gephi were used for statistical analysis,including frequency distribution analysis,association rule analysis,cluster analysis,and complex network analysis.The research results are then compared with previous literature and knowledge graphs,and the clinical diagnosis and treatment rules of chronic gastritis are summarized,and the results are added to the knowledge base.Results1 Research on the construction of a traditional Chinese medicine diagnosis and treatment knowledge graph for chronic gastritis based on clinical case records.1.1 Ontology layer of the Traditional Chinese Medicine knowledge graph for chronic gastritis.The ontology layer of chronic gastritis was constructed by referring to existing literature,standards,and terminology systems,and combining guidance from domain experts.It involves seven entity types,including patients,disease names(including Western medical diseases and Traditional Chinese Medicine diseases),syndromes,symptom signs,formulae,Chinese herbal medicines,and six semantic relationships,including disease,Traditional Chinese Medicine disease,manifestation expression,treatment,usage,and inclusion.1.2 Pattern layer of the Traditional Chinese Medicine knowledge graph for chronic gastritis.Real-world clinical case records of chronic gastritis were used as the data source for knowledge extraction and fusion.The Neo4j graph database was used for storage and display.It was built under seven categories of labels and six relationships,with a total of 4,721 nodes and 90,500 relationships.2 Research on knowledge discovery of TCM diagnosis and treatment of chronic gastritis.2.1 General information:This study included a total of 1,299 patients,of which 653 were male(50.27%)and 646 were female(49.73%),with an average age of 59.05 years.2.2 Diagnostic characteristics:Among the 1,299 patients,a total of 22 Traditional Chinese Medicine syndromes were obtained,with a total frequency of 1,673 times.The five most common syndromes were damp-heat toxin syndrome(301 times,17.99%),damp-heat and blood stasis syndrome(266 times,15.90%),dampness obstruction syndrome(234 times,13.99%),liver-stomach disharmony syndrome(218 times,13.03%),and liver meridian stasis syndrome(99 times,5.92%).A total of 251 symptom sign terminologies were obtained,with a total of 31,537 occurrences.Among them,215 symptom description terminologies were mainly focused on symptoms such as dementia,heartburn,restless sleep at night,dry mouth,belching,acid reflux,and epigastric fullness,and the relationships between these symptoms were relatively close,with strong associations among them.There were 30 tongue diagnosis terminologies,with tongue dark red,tongue red,tongue dark,and tongue pale being the main ones.The coating diagnosis terminologies were mainly yellow coating,greasy coating,thin coating,and white coating.There were six pulse diagnosis terminologies,mainly including wiry pulse,slippery pulse,thin pulse,and sunken pulse,with wiry pulse and slippery pulse being the most prominent.2.3 Treatment characteristics:A total of 1299 patients were treated with 393 different types of Chinese herbal medicines,with a total frequency of 85327 doses.The most commonly used herbs were Huang Qin(Scutellarin baicalensis)with 1812 doses(2.12%),Bai Shao(Paeonia lactiflora)with 1779 doses(2.08%),Huang Lian(Coptis chinensis)with 1774 doses(2.08%),Fu Ling(Poria cocos)with 1750 doses(2.05%),and Ji Nei Jin(Gallus gallus domesticus)with 1683 doses(1.97%).The highest support rates were observed for the herb pairs Dang Gui(Angelica sinensis)-Bai Shao,Bai Shao-Fu Ling,and Huang Qin-Ji Nei Jin,while the herb pairs with the highest confidence levels were Bai He(Lilium brownii)-Wu Yao(Lindera strychnifolia),Fu Ling-Bai Zhu(Atractylodes macrocephala),and Bai Shao-Dang Gui.ConclusionKnowledge graph technology is a powerful tool to organize,display and query data.Based on real-world clinical record data combined with ontology technology,this study constructed the TCM diagnosis and treatment knowledge map of chronic gastritis,including 7 entity concepts such as chronic gastritis patients,diseases,syndromes,symptoms,prescriptions and traditional Chinese medicine,as well as 6 relationships such as illness,TCM disease,phenomenon expression,treatment,use and inclusion,and realized the visualization of the diagnosis and treatment knowledge system of chronic gastritis.Convenient knowledge organization,management and reasoning.By searching the knowledge base,disease related information can be queried,which provides tools and methods for the formulation of clinical diagnosis and treatment plan for chronic gastritis and the study of plan optimization.By supplementing and updating the data in the knowledge map,the most cutting-edge diagnosis and treatment knowledge can be shared,and the diagnosis and treatment ideas and experiences of other physicians can be learned for reference,so as to realize data sharing and exchange.This study made use of the existing medical record data of patients with chronic gastritis,used data mining technology to discover the diagnosis and treatment knowledge of chronic gastritis,and found that age is an important factor affecting chronic gastritis.The core of chronic gastritis lies in the spleen and stomach,but it is also closely related to the functions of the liver,kidney and other viscera.The disease is mixed with deficiency and accumulation,and phlegm and blood stasis are the common pathological products.The main clinical manifestations of the patients are numbness,heartburn,sleepless restlessness,dry mouth,belching,acid regurgitation,and abdominal distension and fullness,etc.In treatment,drugs with functions of invigorating the spleen,dispelling dampness and promoting blood circulation are mainly used,and adjustment is made according to the actual situation of the patients.It can be seen that the pathogenesis of chronic gastritis is complex,the influencing factors are numerous,and the choice of drugs is diverse.By learning the diagnosis and treatment knowledge of chronic gastritis and using the knowledge atlas for query and evidence,the diagnosis and treatment of chronic gastritis can be more accurate,and at the same time,data support is provided for the improvement and optimization of the atlas,and the content and depth of the atlas are enriched. |
PDF Full Text Request