Influence Factor Of Diagnosis And Treatment For Imported Malaria Cases In Henan, 2012-2017: Decision Tree Analysis

ObjectiveImported malaria has been important challenge of China.Typically,fatality rate of malaria increased in China,particularly in Henan Province.This study aims to investigate the influence factors of misdiagnosis and complications of imported malaria cases,based on malaria surveillance data in Henan Province from 2012 to2017.MethodsObject of this study was the imported malaria cases in Henan province,2012-2017,information of cases was extracted from the Infectious Diseases Information Reporting Management System(IDIRMS)and the Parasite Diseases Information Reporting Management System(PDIRMS).The variables in the data containing age,gender,native place,travel history,countries visited,purposes,date of back to China,date and location of onset,date of first attendance,diagnosis and the name of medical institution,similarly the final diagnosis and specific type of complications,etc.,were used to conduct statistical analysis.1.To calculate the statistics included mean,frequency,percentage,median,percentile,etc.2.To draw an onset location map of cases,and present the difference in choice of initial medical institutions by cases from various onset locations and the characteristics in diagnosis and treatment of them.3.To explore the relevant factors of initial diagnosis and complication of cases with difference analysis.4.To generate the tree model with the highest classification accuracy for the dependent variables of initial diagnosis and complications.Results1.This study included 371 imported malaria cases and they were mainly male aging from 20 to 50 years old,containing 319 falciparum malaria cases.2.The proportion of cases living in urban district of Zhengzhou and Xinxiang was49.6% among all the 18 cities with imported malaria cases in Henan.There were63.1% cases in villages and 32.4% cases in urban areas at the time of malaria onset.When they had their first visit to the doctor,the medical institution choices of cases from villages were decentralized,while the choices of cases from urban areas were concentrated on municipal hospitals.3.The medical institutions visited firstly by malaria patients were mainly in provincial(32.3%),municipal(19.9%),and county(15.9%)level.The proportion of misdiagnosed cases in provincial,municipal,county,township,village and individual initial diagnosis institutions were 12.5%,39.2%,54.2%,82.4%,97.8%and 98.2%,respectively,increasing in turn.The proportion of cases with complications in initially misdiagnosed cases was 44.4%,significantly higher than the 18.1% in correct diagnosed cases.4.The variables of category of healthcare facilities for the initial diagnosis,time interval between onset and initial diagnosis were the main factors in the tree model with initial diagnosis as dependent variable and the variable of initial diagnosis became the most significant variable in the tree model of malaria complication.The classification accuracy of the two tree models were 82.2% and74.1%,respectively.Conclusions1.Imported malaria cases in Henan Province showed obvious spatial gathering,with more cases in villages than in urban areas.The spatial distribution of cases from villages was different from that of urban cases.Compared with urban cases,the medical institutions choices for initial diagnosis of cases from villages was scattered.2.The falciparum malaria predominance,delay for doctor visiting by patients and high misdiagnosis rate for malaria by healthcare institutions of lower administrative level were identified as the major factors associated with initial misdiagnosis and complications of imported malaria cases in Henan Province.3.The decision tree model was qualified to identify the characteristics of cases with high rates in misdiagnosis and complication,with good classification accuracy,when applied to small sample data of malaria network reporting with highly correlated independent variables.