In a series of counterfactual simulations, we find that based on the experience in China, preventing sustained community transmissions from taking hold in the first place has the largest impact, followed by restricting population flows from areas with high risks of infections.