Construction Of The ARIMA Prediction Model And Spatial Autocorrelation Analysis Of HFMD Incidence In China

Background Hand,foot and mouth disease is caused by enteroviruses and is an acute infectious disease common in children under 5 years of age.The main symptoms are fever and rash or herpes on the skin,mucous membranes of the hands,feet,mouth,buttocks and throat,with or without symptoms such as nausea,vomiting,sore throat,fatigue,loss of appetite,and mental depression may be accompanied by central nervous system,circulatory,or respiratory complications.In May 2008,China adopted hand-foot-mouth disease as a Class C infectious disease for management.Since 2010,the number of cases of hand-foot-mouth disease often ranks first in China’s statutory report of infectious diseases.HFMD has the largest number of deaths,and the critical illness patients progress rapidly and are prone to death,causing panic to the children’s families.The large-scale epidemic and casualties are social instability factors,which seriously affect children’s health and safety.China has a vast territory and is affected by the complex influence of natural and social factors.The incidence of hand-foot-mouth disease varies greatly in different regions.Due to the imbalance of regional research level and technical standards,different conclusions on the research of the same influencing factors have resulted.At present,most of the epidemiological studies on hand-foot-mouth disease focus on provincial and municipal administrative regions.The research period is mostly 3 to 5 years,and there is still a lack of nationwide and long-term statistical analysis.Objective1.By analyzing the time series of monthly incidence of hand-foot-mouth disease in China from 2008 to 2016,grasping the time distribution of the incidence rate,and exploring the application of the autoregressive moving average model in the prediction of hand-foot-mouth incidence.To provide the the theoretical basis for the prediction of morbidity and scientific basis for the prevention and control of hand-foot-mouth disease.2.This paper establishes a database of hand-foot-mouth disease by region and time,and performs spatial autocorrelation analysis of hand-foot-mouth disease in China,and clarifies the spatial distribution characteristics and agglomeration intensity of the incidence of hand-foot-mouth disease in China,and identifies the key prevention and control area of hand-foot-mouth disease.3.Based on the relationship between regional incidence and meteorological factors under different spatial clustering models,explore the impact of meteorological factors on the incidence.Method1.This article obtains the incidence of hand-foot-mouth disease through the website of the Public Health Science Data Center(http://www.phsciencedata.cn)and the Chinese Center of Disease Control and Prevention(http://www.chinacdc.cn),and uses Excel 2007 to compile the annual and monthly incidence database of 31 provincial administrative regions in China from 2008 to 2016,and 2008-2018 national monthly incidence database of hand-foot-mouth diseases.Obtain and organize 31 provincial-level administrative regions’ meteorological databases through the monthly data set of climate data of China Ground International Exchange Station of China Meteorological Data Network(http://data.cma.cn).Describe it epidemiologically.2.Use SAS 9.1.The monthly incidence data of hand-foot-mouth disease from 2008 to2018 were divided into prediction set and validation set,and ARIMA prediction model of hand-foot-mouth disease incidence was established to predict the incidence of 2019.The model was fitted with validation set,and the effects were evaluated using indicators such as average absolute error(MAE),average absolute percentage error(MAPE)and root mean square error(RMSE).Then predict the incidence rate of 2019.3.Use Arc GIS 10.6 to correlate the incidence data with the national administrative division vector map,generate the incidence geographic information database,and make the thematic map of the incidence,showing the grade distribution of the annual incidence data of each provincial administrative district and the changing state of the incidence of hand,foot and mouth disease with the year.According to the geographical adjacency relation of administrative division in our country,the spatial weight matrix is generated by using the spatial conceptualization method based on inverse distance,and the spatial autocorrelation analysis of incidence is carried out,including global autocorrelation and local autocorrelation4.Use SPSS 20.0 to perform spearman correlation analysis,and analyze the relationship between meteorological factors(mean air pressure,mean temperature,mean relative humidity,and sunshine hours)and incidence in different aggregation modes.Result1.From 2008 to 2018,the total number of reported cases of hand,foot and mouth disease in China was 20 537 199,with an average annual incidence of 137.53 per 100,000 and3667 deaths.Among them,the highest incidence in 2014 was 205.05/100,000,and the lowest incidence in 2008 was 37.00/100,000.The highest number of deaths occurred in2010,with 905 cases,and the lowest number of deaths occurred in 2008,with 126 cases.The overall incidence rate is 12 months cycle.The annual incidence rate has obvious high and low difference,while the incidence peak is mostly concentrated in May-July and the secondary peak is mostly concentrated in September-November.There is a trend of annual growth.2.The prediction model of the incidence of hand,foot and mouth disease in China is ARIMA((12),2,0).The model AIC and SBC values are 604.2607 and 606.9336,respectively,and the residual sequence is white noise.The ARIMA((12),2,0)model was used to predict the monthly incidence of hand,foot and mouth disease in China from January to December 2018.Compared with the validation set,the validation set was within the confidence interval,and the overall basic fit,The average absolute error is 1.467,the average absolute percentage error is 12.72%,and the root mean square error is 2.4335.3.The prevalence of hand,foot and mouth disease varies greatly across provinces in China.The area with the highest average annual incidence(Hainan,417.49/100,000)has the lowest annual incidence(Xinjiang,30.55/100,000)13.7 times.There is spatial heterogeneity in incidence: Guangdong,Guangxi,Hainan,Hunan shows a high-high aggregation pattern;Qinghai,Tibet,Xinjiang,Ningxia shows a low-low aggregation pattern for five years,which is low incidence aggregation area.Jiangxi,Guizhou are often in low-high aggregation state,and in 2009 Guangxi and 2013 Shaanxi show high-low aggregation mode.4.In the four provinces with different spatial aggregation patterns,the monthly incidence change trend is more similar to the monthly change trend of mean temperature,average relative humidity and sunshine hours,and there is no specific rule compared with the change trend of average air pressure.In general,the average temperature,average relative humidity,and hours of sunshine are positively related to the incidence of hand-foot-mouth disease.The impact of average pressure on the disease varies greatly from region to region.Different meteorological factors have different lag effects on the incidence of foot and mouth disease.Conclusion1.The incidence of hand,foot and mouth disease in China has shown an increasing trend every other year,an increasing trend in even years,and a seasonal trend with a 12-month cycle.The peak of incidence occurs in summer and autumn.2.The autoregressive moving average model(ARIMA)can be better applied to the prediction of the monthly incidence of hand-foot-mouth disease.3.Spatial distribution shows obvious north-south differences,and generally shows an upward trend from west to east and north to south.The incidence is lower in the northeast and northwest,and the incidence is higher in the southeast and central provinces.4.Mean temperature,average relative humidity and sunshine hours are the risk factors of HFMD.