Establishment Of Computer Identificationaided System For Differential Diagnosis And Clinical Diagnosis Of Emerging Infectious Diseases

Emerging Infectious Disease(EID) is characterized by strong infectiousness, high propagation speed and high mortality.The transnational communication of EIDs becomes dangerous because of economic globalization and the impact of global climate warming.At the same time,some EIDs have more suddenness and concealment when they are used as biological warfare agents.Medical staffs lack of effective precaution mechanisms due to the lower incidence and the less knowledge of EIDs.So an assistant diagnostic system for basic-level staffis demanded. The study aims at constructing an assistant diagnostic system for differential diagnosis of emerging infectious diseases(viral encephalitis and hemorrhagic fever) so as to lay the foundation for the construction of diagnosis and differential diagnosis of EIDs or even infectious diseases.Methods:1) to collect the data of EIDs and differential diagnosis related diseases and to construct a knowledge base;2) to extract the clinical features of EIDs effectively and to establish digitizing diagnostic evaluation system which could be identified by computer;3) embedded the system to the recording genealogy of clinical data,integrated with knowledge base,provided expert estimation, feature selection and core algorithm including Bayesian network for it;4) to select database frame and to establish a total structured clinical data recording system by using J2EE,relational database technology,structured descriptive components and standardize medical terms.5) the algorithm amendment of system by using cluster analysis,principal component analysis,discriminating information entropy and Bayesian network;6) to integrate the 3 components,establish the system,accomplish trial operation and preliminary identification by using J2EE,R language; 7) to collect cases,verify the system by the retrospective and prospective research.Results:1) completed the collection of relevant diseases' data and established inquiry knowledge base;2) compiled a lot of terms relevant to diagnose,constructed the index system of clinical diagnosis;3) established mathematical evaluation system by using multivariate statistical analysis,information entropy theory and Bayesian network;4) constructed clinical information management system by using structured descriptive components and standardize medical terms; 5) developed a computer-assisted expert intelligent system of EIDs;6) the diagnose accordance rate of retrospective research(314 cases) and prospective research (186 cases)are:61.9%and 59.7%(the first diagnosis),78.1%and 77.9%(the first three),86.6%and 85.4%(the first five) respectively.Conclusion:The system accords with clinical thinking and has functions of suggesting,guiding and training.It is characterized by accuracy,stability,flexibility, compatibility and practicality,which could connect to national epidemic network directly and perfect our clinical monitoring and precautious of infectious diseases.