Authors: Bo Xu*, Department of Earth System Science, Tsinghua University, Jun Cai, Department of Earth System Science, Tsinghua University, Huaiyu Tian, College of Global Change and Earth System Science, Beijing Normal University, Bing Xu, Department of Earth System Science, Tsinghua University
Topics: Medical and Health Geography, Geography and Urban Health, Spatial Analysis & Modeling
Keywords: 2009 influenza A (H1N1), highway network, gravity model, spatial-temporal pattern, node centrality, spatial autoregressive model
Session Type: Paper
Start / End Time: 10:00 AM / 11:40 AM
Room: Bayside C, Sheraton, 4th Floor
Presentation File: No File Uploaded
Influenza A (H1N1) in 2009 caused large amount of cases in mainland China. The epidemics started in several cities with imported cases and spread nationwide as people travelled in traffic transportation systems. Road traffic is the main transport mode in China and conveyed 93% of passengers in 2009. It is important to investigate the effect of road transport and other socioeconomic factors on the spread of 2009 influenza A (H1N1) virus along highway network. We construct a national highway network with 273 prefecture-level cities in mainland China and extract three epidemic related variables (prevalence rate, start week, and duration of epidemics) and thirteen socioeconomic factors (concerning population, income level, medical condition and school education) of each city. Based on the network, we calculate four centrality (degree, betweenness, closeness and eigenvector centrality) of each node and propose two new centrality measures: SumRatio, which is a synthetic indicator to measure the integrated importance of a city to all other cities, and Multicenter Distance, which is the shortest distance from a certain city to one of the eight center cities (Chengdu, Jinan, Beijing, Guangzhou, Shanghai, Fuzhou, Wenzhou and Changsha city), where the epidemics began before May 22, 2009. We find that both of the new metrics perform better than the former four, when conducting correlation analysis between each of them and epidemic variables. In addition, we apply spatial autoregressive models to the data and reveal several statistically significant and quantitative relations between epidemic characteristics and socioeconomic factors, as well as several centrality metrics.