Table 1.

Parameter values and assumptions. Local transmissibility of region i is given by ${\beta }_i={\hat{R}}_0\left(1+ϵ\right)h/D_{\text{I}}$, where ${\hat{R}}_0$is the basic reproductive number assumed when constructing the discretionary policies.

parameter	base case assumptions	sensitivity	sources
$\left\{z_i^{\text{S}}\right\}$, reduction in susceptibility after vaccination, and {pi}, the probability distribution of $\left\{z_i^{\text{S}}\right\}$	vaccine efficacy for susceptibility VES=0.69 when antigenic match is perfect	driven by μ: $z_i^{\text{S}}\leftarrow \mu ·z_i^{\text{S}}$	Riley et al. (2007)
VEI, vaccine efficacy for infectiousness	VEI=0.8 when antigenic match is perfect	driven by μ: $V{E}_{\text{I}}\leftarrow \mu ·V{E}_{\text{I}}$	Longini et al. (2004); Patel et al. (2005)
ϵ, error in estimate of transmissibility	no error, i.e. ϵ=0	[−0.5, 0.5]	Mills et al. (2004); Viboud et al. (2006)
hi, relative hazard for transmissibility in region i	transmissibility is the same for all regions, i.e. hi=1 for all i	[0.5, 1.5]	Mills et al. (2004); Viboud et al. (2006)
δ, relative level of inter-regional mixing	inter-regional mixing is at normal level; δ=1 by definition	[0, 2] $m_{ij}\leftarrow \delta m_{ij},i\ne j$
μ, antigenic match between vaccine strain and pandemic strain	imperfect match with μ=0.7	[0, 1]