Countries should aim to lower the reproduction number R close to 1.0 for the short-term mitigation of COVID-19 outbreaks

Michael E. Hochberg

Note too that sensitivity to start date increases as the time horizon decreases, but this effect is most appreciable for intermediate reductions in ℛ0 (e.g., cf. 20% to 40% reduction in Figs. 2A,B) . In other words, whereas achieving 1.5≲ ℛC<ℛ0 -if commenced early -can result in a substantial reduction in total infections by day 80 (Fig. 2A) , the effect is largely erased on the longer time horizon of 200 days (Fig. 2B) . Interestingly, small delays in starting mitigation measures from the start of a simulation may make little difference to the total number eventually infected. This effect is most noticeable between start days 10 and 30, as mitigation exceeds a 60% or greater decrease in ℛ0 (i.e., ℛC≲1.0), and for the longer time window of 200 days. Note that the lack of sensitivity between start days for lower mitigation in Fig. 2B can be explained by the fact that the epidemic is virtually complete by 200 (see Fig. 1 ). These results thus underscore how attaining mitigation objectives is contingent on the time frame, with rapid responses having short term benefits for a range of mitigation levels ( Fig. 2A) , whereas they have disproportionately higher benefits in the longer-term only for sufficiently pronounced mitigation levels (Fig. 2B) .