In the event that transmission gathers momentum in the source region, our calculations indicate that border screening is likely to delay an Australian epidemic only by several days (rather than several weeks or a few months), unless the number of travelers from the source region entering Australia is reduced to virtually zero or quarantining of travelers arriving from the source region is virtually 100% effective. Our calculations assume that we can achieve the same R in Australia as is achieved in the source region with assistance from the international community. The main drivers of the results on the likely delay are
i. the proportion of recently infected travelers who are pre-symptomatic, and
ii. the exponential growth of infections in the source region (that typically occurs in epidemics).
The results on the likely delay are quite robust to plausible alternative assumptions, unless we can achieve and sustain R<1 in Australia. In the latter event a local epidemic would fade out, repeatedly when there are multiple importations. The exponential growth of infections in the source region (when R>1) means that after a while the mean number of recently-infected individuals who arrive during the pre-symptomatic stage of their infection increases rapidly and any feasible level of border control is unlikely keep all infected individuals out for long.
Border control is only useful for preventing disease entry, and provides very little further benefit once an epidemic has gathered widespread momentum within Australia. Once there is widespread transmission within Australia, its exponential growth will contribute far more cases than will transmission chains from new importations.