Application Research Of Intelligent Medical Record Management Based On Knowledge Graph

With the development of information technology,people’s lives become increasingly inseparable from computers,and computer technology has been applied in all aspects of life.In the medical system,medical record data is an important information for disease diagnosis,treatment and research.As paper medical records are difficult to store,write,query and comparative research,electronic medical record management has received more and more attention in the medical system and has been widely used.After nearly two decades of rapid development,the informationization construction of hospitals in China has begun to rise.Although hospitals have their own Hospital Information System(HIS),how to effectively use electronic medical records data and complete the automatic intelligent diagnosis and treatment of diseases has become an important issue.As an important medium to record patient’s course of disease,examination,operation and doctor’s advice,electronic medical record is helpful for patients’ diagnosis and treatment.At present,each high-level hospital has a mature model for hospital medical record management,but there is no effective methods for crosshospital medical record management.In the application of cross-hospital situation,when patients are transferred to other hospital,many difficulties would be encountered because of the inconsistency of the management standards of the electronic medical records in the hospital information system,for example: a)different electronic medical records systems have inconsistent appellations for the same entity,b)electronic medical records contain a large amount of information,but different hospitals cannot import directly because of the inconsistency of the data structure of the hospital information system,so manual recording is adopted and that would waste a lot of time.In view of above problems,the main problem of this theses is to realize the intelligent analysis and management of electronic medical record,reduce the cost of electronic medical record management,save medical resources and improve the efficiency of hospital operation under the cross-hospital electronic medical record interaction scenario.In order to solve the problem of research,this theses uses a combination of knowledge graph technology and text analysis technology to build a disease knowledge base and complete the intelligent management of electronic medical records.This theses completed the following research contents and achieved corresponding results:(1)This theses analyzes and compares the results of knowledge extraction using Conditional Random Field and Bi LSTM+CRF model in the process of constructing disease knowledge graph.For Chinese disease knowledge,the accuracy rate,recall rate and F1-Measure value of entities extracted by Bi LSTM+CRF model are much higher than Conditional Random Field model.Bi LSTM+CRF model is the priority for building Chinese disease knowledge graph.A complete construction process of medical knowledge graph has been designed according to mentioned demand,which mainly includes five kinds of node information such as disease name,symptom,check,treatment and drug information,as well as eight entity relationships: [disease name,cause,symptom],[disease name,need,check],[disease name,need,treatment],[disease name,commonly used medicine,drug information],[check,find,disease name],[check,find,symptom],[treatment,improvement,symptom],[treatment,need,drug information].(2)According to above problems,a method is designed to automatically analyze unstructured text electronic medical records data into structured course related data and store it in database.When analyzing the medical record data,the method of automatically matching and analyzing the patient’s disease and intelligently recommending the relevant diagnosis and treatment scheme is designed to improve the data utilization efficiency.(3)By using the method of software engineering,this theses describes the design and implementation of the medical record intelligent management system in detail from the aspects of requirement analysis,outline design,detailed design,system implementation and test,and completed the simulation system construction.It can complete the query of the required disease entity and related knowledge.It could also automatically identify the disease entity after inputting the text electronic medical record.It can intelligently give medical advice according to the disease information identified in the medical record.At last,it can automatically extract the unstructured medical record text data into structured data and store it in the database.This theses uses disease knowledge graph to analyze the text information of Chinese electronic medical record,extracts the relevant information of patient’s disease treatment,matches the existing knowledge of knowledge graph,completes the intelligent recommendation of medical treatment and the intelligent management of medical record,and realizes the method innovation of medical record management technology.