Dengue Fever (DF) is a mosquito vector transmitted flavivirus and a reemerging global public health threat. Although several studies have addressed the relation between climatic and environmental factors and the epidemiology of DF, or looked at purely spatial or time series analysis, this article presents a joint spatio-temporal epidemiological analysis. Our approach accounts for both temporal and spatial autocorrelation in DF incidence and the effect of temperatures and precipitation by using a hierarchical Bayesian approach. We fitted several space-time areal models to predict relative risk at the municipality level and for each month from 1990 to 2014. Model selection was performed according to several criteria: the preferred models detected significant effects for temperature at time lags of up to four months and for precipitation up to three months. A boundary detection analysis is incorporated in the modeling approach, and it was successful in detecting municipalities with historically anomalous risk.
Citation Information
Publication Year | 2020 |
---|---|
Title | Spatiotemporal modeling of dengue fever risk in Puerto Rico |
DOI | 10.1016/j.sste.2020.100375 |
Authors | Gavino Puggioni, Jannelle Couret, Emily Serman, Ali S Akanda, Howard S. Ginsberg |
Publication Type | Article |
Publication Subtype | Journal Article |
Series Title | Spatial and Spatio-temporal Epidemiology |
Index ID | 70212845 |
Record Source | USGS Publications Warehouse |
USGS Organization | Patuxent Wildlife Research Center |