Application of Bayesian geostatistical modeling for the assessment of risk for child mortality during the 2008 earthquake in Wenchuan, People’s Republic of China

摘要

A Bayesian geostatistical model was used to identify factors related to topographical variation in the analysis of mortality risk for children less than 5 years of age in the 2008 Wenchuan earthquake in the People’s Republic of China. Epidemiological data from a standardized survey were available from 115 locations across the study area. Physical and demographic factors measured directly at the township level were examined with respect to risk. The geostatistical model explicitly accounted for spatial correlation present in child mortality by fitting a Gaussian linear model to the data. Results identified the role of several factors in explaining geographical heterogeneity in child mortality and show that it was significantly correlated with earthquake intensity, population density, migrant labour, and the percentage of collapsed houses.