Figure 6
Sensitivity analysis considering uncertainty of parameter values. Variability in epidemic curves (large plot) and the distribution of the cumulative number of outpatients (inset), originating from the uncertainty in four parameters (right panel). Parameter values are based on the InfluSim standard configuration [15] with R 0 = 2.5, except those listed at the end of this legend and indicated by superscripts1. The sensitivity analysis extends the scenario shown in Figure 5, where antivirals are available for 10% of the population and are distributed from day zero2, and where contact reduction measures3, including the isolation of cases4, are initiated three weeks after the introduction of infection (scenario "day 21"). Right panel: parameter values for each realization are sampled independently from normal distributions as shown (means given in bold, 99% of the values lie within the range specified by dotted lines, except b A which is truncated). R 0: basic reproduction number, x 50: cumulative infectivity during the first half of the symptomatic period, b A: relative infectivity of asymptomatic cases, f c: antiviral treatment reduces infectivity by a factor of 1-f c. For each parameter, an increase of the value aggravates the epidemic. Large plot: from a hundred random realizations, we selected the two most extreme epidemics, and eight epidemics homogeneously placed between them. The epidemic with N 0 = 20800 is caused by parameter values drawn from the left tail of the corresponding distributions, and the epidemic with N 0 = 5000 is caused by parameter values drawn from the right tail of the corresponding distributions (see right panel). The epidemic curves show a plateau or a second wave when antiviral stockpiles are exhausted while the proportion of susceptibles is still large enough to allow for further propagation of infectives (thin curves in black); for optimistic parameter combinations (e.g. small R 0), the available stockpiles last over the whole period of the intervention and the epidemic curve proceeds without a plateau (bold curves in grey). Inset: distribution of cumulative number of outpatients obtained from 1,000 random realizations.1:Parameter modifications are given in the following and terms in italics refer to terms in the InfluSim user interface. 2: Antiviral availability: 10%. Treatment fraction: 100%, for both, Treatment of severe cases and Treatment of extremely sick cases. 3: Contact reduction by: 20%. Range of days: day 21–360. 4: Moderately sick cases: 10%, Severe cases (home): 20%, Severe cases (hospital): 30%. Range of days: day 21–360.