The Epidemiological Characteristics Of Epidemic Hemorrhagic Fever In Liaoning Province And Its Prediction Based On Meteorological Factors

Background At present,global warming has become one of the major environmental problems.The impacts of climate change on human infectious diseases are mainly manifested in a series of changes,such as the survival and variation of pathogens,the change of the active regions of vector animals,and the emergence of new pathogens and new infectious diseases caused by them.Among infectious diseases which are greatly affected by climate change,especially vector-borne diseases.Epidemic hemorrhagic fever(EHF),hereinafter referred to as EHF,is also known as hemorrhagic fever with renal syndrome because of its three main characteristics of fever,hemorrhage and kidney damage.It is mainly a multi-host natural epidemic disease caused by Hantavirus.Since the discovery of EHF,the number of cases has increased year by year,and has expanded from local areas to the whole country.According to the analysis of monitoring data,EHF cases have been reported in all provinces in mainland China.In 2005,China CDC set up national surveillance sentinels in 40 key districts and counties of 22 provinces,and carried out EHF case surveillance systematically and normatively throughout the country.With the improvement of people’s living standards,the demand for health is also increasing,which poses a new challenge to the traditional methods of disease prevention and control.It is necessary to predict diseases in advance and to prevent and control diseases according to the predicted results.On the basis of previous studies,this study intends to analyze the epidemic characteristics of EHF in Liaoning Province in time and space,clarify the lag impact of temperature on the incidence of EHF,and build a prediction model on this basis,so as to provide a basis for rational allocation of health resources and precise prevention and control of diseases.Objective To master the three distribution characteristics of epidemic hemorrhagic fever in Liaoning Province.The hysteresis effect of temperature on EHF onset time in Dandong,Huludao and Shenyang were analyzed.The distribution lag non-linear model and generalized additive predictive model of EHF onset time and meteorological factors were constructed to predict the number of EHF cases in different months,and to provide theoretical basis for the early formulation of targeted EHF prevention and control measures.Method From 2005 to 2017,EHF case information was collected from the"China Center for Disease Control and Prevention-China Legal Infectious Disease Monitoring System".From 2005 to 2017,the meteorological data were downloaded from the National Meteorological Science Data Sharing Service Platform,which is an open free website,then EHF case database and monthly and weekly database of meteorological factors databases were established.The epidemic characteristics of EHF cases,such as gender,age and incidence area,were analyzed descriptively by SPSS 19.0.Evolution of symptoms and statistical inference of differences were made.Distribution characteristics of EHF cases were analyzed by ArcGIS software.The relationship between onset time of EHF cases and temperature was analyzed by distributed lag non-linear model(DLNM).The time prediction model of EHF cases was constructed by Quasi-Poisson generalized additive model(GAM).P<0.05 is the difference with statistical significance.Result From 2005 to 2017,17335 cases of EHF were reported in Liaoning Province.The number of cases decreased first and then increased.The incidence rate ranged from 2.193/100 000 to 9.9931/100 000.Most of the EHF cases were farmers(61.55%,10670/17335).The occupational distribution of cases in different years had statistical significance(χ2=93.214,P=0.000).The average age of cases in each year was 40.04 to 48.02 years old.80%(13869/17335)of EHF cases occurred from October to December and from March to May of the following year,forming the autumn and spring peaks of EHF respectively.There was significant difference in the monthly distribution of EHF cases in different years(χ2 = 822.287,P = 0.000).For Dandong City,Huludao City and Shenyang City,DLNM was used to analyze the lag effect of temperature on the number of EHF cases per week.The effect of low temperature on the incidence of EHF cases in Dandong City was more significant than that of high temperature;the effect of temperature on the EHF epidemic in Huludao City reached the greatest level at-5℃;and the lag time of the high temperature in Huludao City and Shenyang City can reach 20 weeks.The hysteresis effect of temperature in the three places is shown as that when the temperature is lower(5th and 25th quantile of the average weekly temperature),the effect of temperature immediately occurs and reaches its maximum at lagO,while when the temperature is higher(75th and 95trh quantile of the average weekly temperature),the lag effect of temperature is very long,Dandong City,Huludao City and Shenyang City reached the highest cumulative risk(CRR)at lagO(CRR = 3.54,95%CI:1.71-4.35),(CRR = 1.29,95%CI:1.02-1.45),(CRR = 2.21,95%CI:1.04-3.72),at-8.30(5 th),at-8.79(5 th)and-4.44(25 th),respectively,while the three cities at high temperature of 24.23℃(95th),and at 24.61℃(95th)and 25.28℃(95th),the cumulative risk reached the maximum at lag20(CRR=1.35,95%CI:1.00-12.79),(CRR = 1.97,95%CI:1.24-16.53)and(CRR = 2.43,95%CI:1.11-4.37).Spearman analysis was used to adjust the variables included in the model and their lagging order.Variables affecting EHF in that month were selected,including temperature,relative humidity,rainfall and the number of EHF cases last month.On the basis of controlling the long-term trend,a Quasi-Poisson generalized additive model was constructed.The model fitted well and could predict the number of EHF cases one month earlier.Conclusion The number of EHF cases in Liaoning Province have increased in recent years.The epidemic patterns in different areas of Liaoning Province are different.The gender distribution,age distribution and occupational distribution of EHF cases have statistical significance.The lag effect of temperature in the study area is that the effect of low temperature occurs rapidly and the effect of high temperature lags for a long time.GAM model use EHF and meteorological factors to construct a predictive model.