不同人口流动模式下城市传染病时空传播模型适用性研究  

Research on the Applicability of Urban Infectious Disease Spatiotemporal Transmission Model Under Different Population Flow Patterns

作　　者：薛山 廖一兰[1] 李春林 胡艺[3,4,5] XUE Shan;LIAO Yilan;LI Chunlin;HU Yi(State Key Laboratory of Resources&Environmental Information System,Institute of Geographic Sciences&Natural Resources Research,Chinese Academy of Sciences,Beijing 100101,China;College of Resources and Environment,University of Chinese Academy of Sciences,Beijing 100049,China;Department of Epidemiology and Biostatistics,School of Public Health,Fudan University,Shanghai 200032,China;Key Laboratory of Public Health Safety,Ministry of Education,Shanghai 200032,China;Laboratory for Spatial Analysis and Modeling,School of Public Health,Fudan University,Shanghai 200032,China)

机构地区：[1]中国科学院地理科学与资源研究所资源与环境信息系统国家重点实验室,北京100101 [2]中国科学院大学资源与环境学院,北京100049 [3]复旦大学公共卫生学院流行病学与卫生统计学系,上海200032 [4]教育部公共卫生安全重点实验室,上海200032 [5]复旦大学公共卫生学院空间分析与建模实验室,上海200032

出　　处：《地球信息科学学报》2023年第1期208-222,共15页Journal of Geo-information Science

基　　金：国家自然科学基金项目(42171419)

摘　　要：空间显式传染病传播模型由于能够较好地模拟城市内部的传染病传播过程,在城市疫情防控中得到了广泛应用。但现有的空间显式模型大多按照人口随机流动模式来模拟城市人口流动行为,未考虑城市中以通勤为主的人口流动模式。本文以长沙市的2014年第1~49周手足口病传播过程为例,将4、8、24邻域扩散以及重力模型、辐射模型共5种不同的人口流动模式与周期SEIR动力学模型结合建立起五种传染病传播模型,并通过比较模型时间和空间精度来明确不同人口流动模式应用在城市内传染病时空传播模拟中的各自优势和适用范围。实验结果表明:每种人口流动模式在城市内传染病时空传播模拟中都有其使用前提和适用范围。基于24邻域扩散的传播模型对手足口病病例数拟合精度最高(RMSEspatial_ts=0.58,RMSEspatial_st=0.95),基于重力模型的传播模型对城市内部传染病时空传播模拟中的传播趋势的模拟精度最高(rspatial_ts=0.46,rspatial_st=0.39),且对长沙市人口相对稠密区域(格网常住人口数大于第三四分位数区域)模拟具有显著的精度优势(rspatial_ts=0.55,RMSEspatial_ts=0.92),但在疫情早期对手足口病未来传播的预测能力上较差。基于辐射模型的传播模型作为无参数模型,虽然整体时空精度上表现欠佳,但在人口稠密区域显现出较好的拟合效果(rspatial_ts=0.50,RMSEspatial_ts=1.0),在五种模型中具有最佳的手足口病传播预测能力。The spatially explicit infectious disease transmission model has been widely used in urban epidemic prevention and control since it can better simulate the transmission processes of infectious diseases in cities.However,most of the existing spatially explicit models simulate the flow behavior of urban population based on the random population flow pattern,and do not consider the commuting-based population flow pattern in the city.Taking the Hand,Foot,and Mouth Disease(HFMD)transmission process in Changsha city from the 1st to the 49th week of 2014 as an example,this paper combines periodic SEIR dynamic models with five different population flow models including 4,8,24 adjacent diffusion models,gravity model,and radiation model.Thus,five infectious disease transmission models are formed,and the respective advantages and applicable scope of different transmission models are clarified by comparing their temporal and spatial accuracy.The experimental results show that each model has its premise and scope of application.The transmission model based on 24-neibor diffusion has the highest fitting accuracy of the number of HFMD cases(RMSEspatial_ts=0.58,RMSEspatial_st=0.95).The model based on gravity model achieves the highest accuracy in simulating the transmission trend of HFMD in Changsha(rspatial_ts=0.46,rspatial_st=0.39),with even higher simulation accuracy in the relatively densely populated areas in Changsha(the grid resident population is greater than the third quartile area)(rspatial_ts=0.55,RMSEspatial_ts=0.92).But its prediction ability of HFMD transmission in the early stage of HFMD epidemic is poor.The transmission model based on the radiation model is a parameter-free model,although its overall spatial-temporal accuracy is not high,it shows a good fitting effect in densely populated areas(rspatial_ts=0.50,RMSEspatial_ts=1.04)and has the best HFMD transmission prediction ability among the five models.

关 键 词：空间显式传染病传播模型 人口流动模式 邻域扩散 重力模型 辐射模型 城市内传染病传播模拟 长沙市 

分 类 号：O175[理学—数学] R181.2[理学—基础数学]