Fig 1.
The observed yellow fever outbreak in Luanda from December 5, 2015 to August 18, 2016.
Green dots denote the sum of probable and confirmed cases; black dots denote confirmed cases; red bars denote the confirmed death; and blue dots denote vaccine coverage. The vertical grey dashed line denotes the time point when the YF vaccination campaign was initiated.
Fig 2.
Flowchart of the yellow fever model.
Black arrows represent infection status transition paths, red dashed arrows represent transmission paths and the blue arrow represents the vaccination pathway. Square compartments represent host classes and circular compartments represent vector classes. Red compartments represent infectious classes, and gray compartments are the simulated weekly reported cases (Zh) and deaths (Yh).
Table 1.
Summary of parameters.
Table 2.
Parameter summary for two scenarios.
X.0 denotes X(t = 0), which is the number individuals in X class at the beginning of the study period.
Fig 3.
Model fitting results under two scenarios: Scenario 1 (weak infectivity with ψ = 0.1) in panels (a,b) and scenario 2 (strong infectivity with ψ = 0.5) in panels (c,d).
Black line with circles denote reported cases (in (a), in the form of square-root) and reported deaths (in (b), in the form of square-root), and red line denotes model simulation median. Blue dashed line denotes the fitted basic reproduction number, , and the green dashed line shows the calculated host susceptible proportion, S(t) (or Sh(t)/Nh in the model Eq 1). Shaded region represents 95% bound of 1,000 model simulations. Vertical dashed line indicates the start date of the vaccination campaign. Inset panel shows BIC as a function of the number of nodes (nm). The lowest BIC is attained at nm = 7 in both scenarios, which is used in the main panel. Parameter values are listed in Table 2.
Fig 4.
Reconstructed transmission rate Q(t) via Fine & Clakrson’s method (with a three-week shift to the left) and our estimated .
Fig 5.
Simulation results of scenario 1 under three deferred vaccination campaign scenarios: 60 day delay in panels (a,b), 120 day delay in panels (c, d) and 180 day delay in panels (e,f).
Black line with circles denote reported cases (e.g., in (a), in the form of square-root) and reported deaths (in (b), in the form of square-root), red line denotes model simulation median and blue dashed line is the fitted basic reproduction number, . Shaded region represents 95% range of 1,000 simulations. The vertical dashed line represents initiation of the vaccination campaign. The number of nodes, nm = 7, is adopted.
Table 3.
Impacts of vaccination campaign delay under scenario 1: Weak infectivity.
Table 4.
Vaccine doses used during yellow fever outbreak (from Dec 5, 2015 to Aug 15, 2016) under four vaccination delay situations.
Fig 6.
The Partial Rank Correlation Coefficients (PRCCs) of basic reproduction number (panel (a)) and total deaths (panel (b)) with respect to model parameters.
Sh.0 denotes the initial susceptible ratio (Sh.0/Nh). The black dots are the estimated correlations and the bars represent the 95% C.I.s. The ranges of parameters are given in Table 1.