The Research On Auxiliary Diagnosis Of Brucellosis Based On Naive Bayesian Model

Objective: Brucellosis is a worldwide zoonotic disease caused by Brucella.Humans usually infect diseases by direct contact with infected animals,eating or drinking contaminated animals and their products,or inhaling contaminated air.In China,at present,we mainly use the following three methods for diagnosis: epidemiological contact history,clinical symptoms,and laboratory examinations.Since the clinical manifestations of human brucellosis are mainly fever,joint pain,fatigue,sweating,etc.,which are non-specific manifestations,the condition was easy to be delayed due to misdiagnosis and missed diagnosis in the early stage of disease may result in worse health issues,making it to be a recurring disease and difficult to cure,which is not only caused serious damage to health,but also brought huge economic losses to people.In order to provide clinicians with reference for the outbreak or clinical diagnosis of brucellosis,and make a diagnosis as soon as possible,this study attempts to explore the incidence,infection pathways,clinical manifestations and existing diagnostic methods of brucellosis in our country.This study collects the clinical characteristics of human brucellosis in the database,apply Naive Bayes theory,establish model,and design a program to provide references for clinicians in the diagnosis of human brucellosis.Methods: The data on the number and incidence of brucellosis,epidemic hemorrhagic fever,epidemic cerebrospinal meningitis,typhoid and paratyphoid fever in China’s mainland from 2005 to 2018,and the number of cases by region were from the China Public Health Science Data Center(http: //www.phsciencedata.cn/).This study described the incidence and trend of human brucellosis in China from 2005 to2018,and collected the information on the clinical manifestations of human brucellosis,epidemic hemorrhagic fever,and epidemic in the three Chinese databases of CNKI(China National Knowledge Infrastructure),Wanfang Data Knowledge Service Platform,China Science and Technology Journal Database from 2009 to 2018.The model was established based on the Naive Bayes theory.In R 4.0.2,the Naive Bayes formula was mainly used to estimate the probability of a disease that is not specific to a certain clinical manifestation,so as to assist in the diagnosis of diseases with similar clinical manifestations and easy to be confused.The basic information of brucellosis reported by Huludao city,Liaoning Province was used to verify and evaluate the model,and visual Fox Pro was used to design and complete a program to present the Naive Bayesian model.Results: 1.From 2005 to 2018,China reported a total of 513,508 new cases of brucellosis,and the number of new cases reported each year exceeded 15,000.In 2014,the number of reported cases reached 57,222.Except for human brucellosis,the incidence of the other three diseases has been generally stable after 2006 and has shown an overall downward trend,but the incidence of human brucellosis has been fluctuating and rising overall.2.Among the 207 articles and 12,308 patients collected from the three databases of CNKI(China National Knowledge Infrastructure),Wan Fang Data Knowledge Service Platform,China Science and Technology Journal Database,10,221 patients had fever,accounting for 83.04%.3.Among the 192 confirmed cases reported on Huludao in 2014,169 reported that they had contact with animals.One of them was in contact with cattle and others were in contact with sheep.The main route of exposure with animals were breeding,grazing,slaughter,and breeding etc.4.Among the 192 confirmed cases of human brucellosis reported in Huludao in 2014,fever accounted for 89.58% of clinical manifestations,hyperhidrosis accounted for 95.31%,muscle and joint pain accounted for 93.75%,fatigue accounted for 92.19%,and hepatomegaly accounted for 5.73%,splenomegaly accounted for 31.25%,lymphadenopathy accounted for 1.56%,and testicular enlargement accounted for 4.17%.5.Among the 192 patients with brucellosis included in the Huludao infectious disease reporting system in 2014,185 patients were correctly identified as brucellosis,and the accuracy of the first identification was96.35%.Conclusion: 1.The incidence of human brucellosis in our country fluctuated continuously from 2005 to 2018,but the overall trend was increasing.2.Fever is the main symptom of human brucellosis,epidemic hemorrhagic fever,epidemic cerebrospinal meningitis,typhoid fever and paratyphoid fever.The four diseases also have other clinical manifestations in common.3.The Naive Bayes model established in this study has a high accuracy and can be an auxiliary diagnosis tool.