Spatiotemporal Analysis of Communicable Diseases based on Stochastic Epidemic Tree: A Case Study of Dengue Fever in Guangzhou City, China

Authors: Meifang Li*, Sun Yat-sen University; Dartmouth College, Xun Shi, Dartmouth College, Xia Li, Sun Yat-sen University; East China Normal University
Topics: Medical and Health Geography, Geographic Information Science and Systems, Spatial Analysis & Modeling
Keywords: epidemic trees, basic reproductive ratio, spatial epidemic models, dengue fever, Guangzhou
Session Type: Paper
Day: 4/11/2018
Start / End Time: 1:20 PM / 3:00 PM
Room: Napoleon C3, Sheraton 3rd Floor
Presentation File: No File Uploaded


Traditionally, epidemic models for characterizing communicable diseases focus on the temporal dimension and are non-spatial. The case-reproduction ratio (Rt), the most widely used measurement for characterizing an epidemic process, is usually estimated as a simple mean value for the entire study area, without considering the spatial variation. In this study, we implemented epidemic tree method to spatialize the characterization of a communicable disease epidemic, and applied it to the 2013 dengue fever epidemic in Guangzhou City, China. Compared with previous studies, we included time as a dimension in calculating the spatiotemporal distance between cases, and allow its weight to be adjustable; the tree was constructed through a stochastic process that builds a parent-child linkage based on the conditional probability. Based on the constructed trees, we calculated Rt at three spatial scales, namely global, tree-specific, and pixel-wise. We further conducted correlation analysis, also at the three scales, between weekly Rt and corresponding climate factors, to detect dengue-climate associations. The findings indicate that Rts at three different spatial scales, and the dengue-climate associations detected based on them, offer different information that is important in epidemiological studies and disease control practices. Through this study, we explored what can be done to spatialize the epidemiological models, what information can be extracted from the output of this spatialization, and then how to use the extracted information. We also point out that spatialization of Rt is the basic process of mapping a communicable disease, corresponding to the spatialization of incidence /prevalence in mapping a chronic disease.

Abstract Information

This abstract is already part of a session. View the session here.

To access contact information login