Fig 1.
Spatial distribution and timeseries of dengue cases and containment activities.
Spatial distribution and bi-weekly aggregate timeseries of dengue cases and containment activities between 2012 and 2017 in Lahore (left), and between 2014 and 2017 in Rawalpindi (right). Timeseries plots for each city show aggregated numbers from all sub-city spatial units. Spatial units colored to show relative distribution of cases and each containment activity within each city; lighter areas with lower numbers (total number of cases and containment activities summarized in Tables A and B in the appendix). Adulticides are color coded orange, larvicides coded red and source reduction activities coded as brown. Spatial units boundaries for each city are shown as black polygons. In Rawalpindi, spatial units are defined only in areas where cases occurred. All figures produced by the authors.
Fig 2.
Spatial dependence of cases in Rawalpindi and Lahore.
Spatial dependence of cases occurring within 30 days of index cases in Rawalpindi and Lahore. The spatial window of the analysis (d2 minus d1) is maintained at 500 m when d2 is greater than 500 m, and observations are made by sliding the window at intervals of 100 m. For d2 less than 500 m, d1 is equal to zero and observations are made by increasing d2 at intervals of 100 m. Spatial dependence estimates are plotted at midpoint of the spatial window. The time window t2 − t1 is set to 30 days. 95% CI is shown as shaded area around estimate.
Fig 3.
Variation in the effect of containment activity versus the distance (in meters) from index cases.
Variation in the effect of containment activity, ξa(d1, d2), versus the distance (in meters) from index cases using data from Rawalpindi and Lahore across various values of matching distance m. Values of ξa are calculated using containment and non-containment cases which appear in an m = 1000 m radius of each other. The spatial window of the analysis (d2 minus d1) is maintained at 500 m when d2 is greater than 500 m, and observations are made by sliding the window at intervals of 100 m. For d2 less than 500 m, d1 is equal to zero and observations are made by increasing d2 at intervals of 100 m. Spatial dependence estimates are plotted at midpoint of the spatial window. Values below 1 show a lower probability of new cases appearing around a case in proximity of a containment activity, compared to a non-containment case. The time window t2 − t1 is set to 30 days. 95% CI from bootstrapping 100 replications are shown as shaded areas around estimates.
Fig 4.
Variation in reproductive number (R0) of dengue, with variation in parameters, in Lahore.
Variation in reproductive number (R0) of dengue, with variation in parameters, in Lahore during 2012–2017 for different incidence reporting rate values. For containment activities, x-axis represents the total number of containment activities performed, in a spatial unit, in a lagged time step and any residue from previous weeks. For temperature, x-axis represents the average temperature in Celsius in a lagged time step. For rainfall, x-axis represents the number of rainfall days in a lagged time step. For population density, x-axis represents individuals per 1 sq.km.
Fig 5.
Variation in reproductive number (R0) of dengue, with variation in parameters, in Rawalpindi.
Variation in reproductive number (R0) of dengue, with variation in parameters, in Rawalpindi during 2014–2017 across various incidence reporting rate values. For containment activities, x-axis represents the total number of containment activities performed, in a spatial unit, in a lagged time step and any residue from previous weeks. For temperature, x-axis represents the average temperature in Celsius in a lagged time step. For rainfall, x-axis represents the number of rainfall days in a lagged time step. For population density, x-axis represents individuals per 1 sq.km.