Figure 1.
The 2003 Cairns epidemic in numbers.
(A) Weekly number of confirmed dengue 2 cases (bars) and of indoor residual insecticide sprays (line) performed in the city of Cairns during January–August 2003, and (B) weekly variation in temperature (mean, minimum and maximum) and total precipitation over the same period. Time is measured in weeks since the onset of symptoms of the introduced case (IC).
Table 1.
Adjusted dengue incidence rates per age class for the city of Cairns during January–August 2003.
Figure 2.
Spatial progression of the dengue epidemic that affected the city of Cairns, Australia, during January–August 2003.
Time is represented as weeks since the onset of symptoms of the introduced case (IC). Red circles represent cases confirmed on the week shown, whereas blue circles represent the cumulative cases until that occurred up to that week. Video S1 is an animated version of this figure.
Figure 3.
Directionality analysis for the location of dengue cases within dmax (the distance up to which clustering of cases around the introduced case (IC) occurred).
(A) Wavelet directionality analysis (WDA) for dengue cases. (B) Standard deviation (SD) ellipses showing the spatial anisotropy of dengue cases. (C) WDA for houses. (D) Wind rose plots showing the directionality (measured in 20 degree bins from north cardinal point) of daily prevailing wind direction and speed at 9:00 AM and 3:00 PM during February–March 2003 (the first 2 months of the epidemic, when most dispersal around the index case occurred).
Figure 4.
Significant space-time clustering (assessed by the Knox test) of dengue cases in the city of Cairns, Australia, during January–August 2003.
Red circles and numbers identify each individual space-time cluster. Detailed information about each cluster can be found in Table 2.
Table 2.
Description of each space-time cluster identified for the dengue epidemic that affected the city of Cairns during January–August 2003.
Figure 5.
Using human infection as a marker of dengue virus epidemic spread.
(A) Association between the proportion of dengue cases that occurred on a given week (colors) and the spatial and temporal distances to a putative index case (PIdC) within a space-time cluster. PIdC can be interpreted as the most likely origin of a space-time cluster. (B) Mean distance from a confirmed dengue case to a PIdC according to the time (in weeks since the onset of PIdC symptoms) such case started presenting dengue symptoms. SD represents standard deviation of the mean value.
Figure 6.
Posterior mean distributions obtained from the semi-parametric Bayesian structured additive regression (STAR) model applied to the weekly dengue infection in Cairns during the 15 weeks post virus introduction.
(A) effect of time (ftime) with 95% credible intervals. (B) mean posterior spatial effect (fspat). (C) effect of IRS (measured as cumulative percentage of premises sprayed around premise i within t0 and t) with 95% credible intervals. (D) mean posterior spatial effect of IRS. White dots in maps represent the location of the index case's (IC) residence.