Risk Assessment of Novel Coronavirus COVID-19 Outbreaks Outside China

Abstract

We developed a computational tool to assess the risks of novel coronavirus outbreaks outside of China. We estimate the dependence of the risk of a major outbreak in a country from imported cases on key parameters such as: (i) the evolution of the cumulative number of cases in mainland China outside the closed areas; (ii) the connectivity of the destination country with China, including baseline travel frequencies, the effect of travel restrictions, and the efficacy of entry screening at destination; and (iii) the efficacy of control measures in the destination country (expressed by the local reproduction number R loc ). We found that in countries with low connectivity to China but with relatively high R loc , the most beneficial control measure to reduce the risk of outbreaks is a further reduction in their importation number either by entry screening or travel restrictions. Countries with high connectivity but low R loc benefit the most from policies that further reduce R loc . Countries in the middle should consider a combination of such policies. Risk assessments were illustrated for selected groups of countries from America, Asia, and Europe. We investigated how their risks depend on those parameters, and how the risk is increasing in time as the number of cases in China is growing.

Keywords: COVID-19; branching process; compartmental model; interventions; novel coronavirus; outbreak; risk assessment; transmission; travel.

Figure 1

Final epidemic sizes in China, outside Hubei, with $R_0=2.1,2.6,3.1$, as a function of the time when the control function $u\left(t\right)$ reaches its maximum (in days after 23 January). Rapid implementation of the control generates much smaller case numbers. The inset shows the estimations of the ascertainment rate for the week 25–31, with average $0.063$, based on the ratio of confirmed cases and the maximum likelihood estimates of the case numbers from exportation.

Figure 2

(Left) Risk of major outbreaks as a function of cumulative number of cases in selected countries, assuming $R_{\mathrm{loc}}=1.6$ and baseline connectivity to China. Other countries in South America, including Mexico, are inside the green shaded area. (Right) The effects of reductions of imported case numbers (either by travel restriction or entry screening) in the USA and Canada, assuming $R_{\mathrm{loc}}=1.4$.

Figure 3

Outbreak risks for highly connected countries in Asia. Thailand and the Republic of Korea are plotted; the curves for Japan and Taiwan are in between them. (Left) We plot the risk vs. the efficacy of prevented importations when the cumulative number of cases reaches 150,000. (Right) C = 600,000. Black parts of the curves represent situations when the four countries are indistinguishable.

Figure 4

Selected European countries with high, medium, and low connectivity to China. (Left) The outbreak risk is plotted assuming their baseline connectivity $\theta$, and $R_{\mathrm{loc}}=1.4$ for each country, as the cumulative number of cases is increasing. A significant reduction in the risks can be observed (Right), where we reduced $R_{\mathrm{loc}}$ to 1.1 and assumed a 50% reduction in importations.

Figure 5

Heatmap of the outbreak risks as functions of $\theta$ and $R_{\mathrm{loc}}$, when C = 200,000. The arrows show the directions corresponding to the largest reductions in the risk.