Figure 4.

Mean and worst case epidemic elimination probabilities. We compute the mean (${\overline{z}}_s$) and worst case (z_s,min) elimination probability curves using the incidence data from the MERS-CoV epidemic in South Korea in 2015 as in [29], i.e. we compute probabilities of sequences of zero future case days based on the available data and samples from the reproduction number distributions of the listed control scenarios (see main text). This procedure is repeated 2000 times to generate elimination curve distributions, from which we extract ${\overline{z}}_s$ and z_s,min. The curves are for various offspring dispersion parameters, k, rising from 0.1 to 2 under a mean controlled reproduction number of ρμ = 0.5. We find that increasing heterogeneity (smaller k) leads to earlier mean elimination times (the ${\overline{z}}_s$ are larger) but later maximum (worst case) elimination times (the z_s,min are smaller).