A Study On Spatio-Temporal Heterogeneity Of The Association Between Apparent Temperature And Hand,Foot And Mouth Disease And Identification Of The Threshold Of Apparent Temperature For Health Warning

BackgroundHand,foot and mouth disease(HFMD)is an important public health problem in China.Meteorological factors are closely related to the incidence of HFMD.Previous studies have paid more attention to the impact of one single meteorological factor on the incidence of HFMD.Comprehensive indexes of different meteorological factors may better reflect the complex relationship between weather and HFMD,but there are few related studies.Objective and significanceTo select appropriate comprehensive meteorological index(apparent temperature,AT),study the association between the index and HFMD,and analyze the spatiotemporal heterogeneity of the association.Then to identify the thresholds of the comprehensive index for HFMD risk warning.MethodsReported cases of HFMD and meteorological data at the provincial and monthly scales of the whole nation,and municipal and daily scales of Guangdong,Anhui and Jilin were collected.The reported cases and meteorological data were used to fit the distribution lag non-linear model(DLNM)of several comprehensive meteorological indexes(apparent temperature,AT)and HFMD,then the most optimal comprehensive index was selected according to the principle of smallest Akaike information criterion values and the residual sum of square in model fitting.The exposure response relationship between the selected comprehensive index and HFMD were evaluated by two-stage method that combined DLNM with Meta-analysis,and the spatial heterogeneity of exposure response relationship of AT-HFMD was analyzed by Metaregression.The thresholds of AT for HFMD risk warning were identified by random forest model.The setting of thresholds was proved to be reasonable by comparing the average daily incidence cases and attributed incidence cases of each risk level.Main results1.AT calculated by Steadman RG method was the most optimal comprehensive meteorological index for this study.2.There was a nonlinear and positive correlation between AT and the occurrence of HFMD.National-level analysis showed that the relationship curve of AT-HFMD increased first and then decreased,with the highest risk at 17.7 ℃(RR=2.61,95%CI:2.00～3.42).The risk of HFMD in Inner Mongolia and Northeast China increased significantly with the increase of AT,while that in North,Central,Southwest and Northwest China decreased after the peak,and that in East,South China and Qinghai-Tibet area increased to a certain height and tended to become stable.In the analysis of the three provinces,higher AT had a greater and longer impact on HFMD compared with lower AT.Cumulative exposure response curves showed that the risk of HFMD for three provinces in a whole increased with the increase of AT.In analysis for each province,the effect curves of Guangdong increased first and then tended to be flat,while that of Anhui increased first and then decreased,and that of Jilin increased monotonously.In analyzing influencing factors of spatial heterogeneity of AT-HFMD relationship,the heterogeneity of the result was the smallest in the regression with latitude and longitude,average temperature and average sunshine hours included.The analysis results at different time scales showed that the temporal heterogeneity of the association between apparent temperature and HFMD was insignificant.3.The risk levels of HFMD in Guangdong and Jilin were identified as low(≦10.9℃ for Guangdong;≦-2.3℃ for Jilin),medium-low(10.9℃～17.9℃;-2.3℃～12.2℃),medium-high(17.9℃～24.4℃;12.2℃～23.4℃)and high(>24.4℃;>23.4℃),that in Anhui were identified as low(≦-3.9℃),medium(-3.9℃～2.0℃)and high(2.0℃～18.9℃)levels for cold effect,and high(18.9℃～31.9℃)and low(>31.9℃)levels for hot effect.Validation showed that daily incidence cases and attributable incidence cases increased with the increase of risk level,and the difference between the incidence cases of each level was statistically significant(P<0.05).ConclusionsAT is a comprehensive index that integrates temperature,relative humidity and wind speed.The increase of AT will increase the risk of HFMD.There is spatial heterogeneity in this relationship,which can be partly explained by differences in geographical regions and weather conditions.In this study,the identified thresholds of AT for HFMD early health warning have been proved to be effective.