Estimating number of global importations of COVID-19 from Wuhan, risk of transmission outside mainland China and COVID-19 introduction index between countries outside mainland China

Haoyang Sun; Borame Lee Dickens; Mark Chen; Alex Richard Cook; Hannah Eleanor Clapham

Next, we propagated the uncertainty in the estimated number of imported COVID-19 cases from Wuhan, and estimated the probability that a local outbreak would occur and sustain for at least three generations (hereinafter referred to as "local outbreak risk") for each country or territory outside mainland China. We modelled the offspring distribution of each case as a negative binomial distribution, with mean equal to the basic reproduction number estimated by Riou et al. 15 , and dispersion parameter assumed to be equal to that of SARS-CoV. 16 Using the first-step analysis, the local outbreak risk can be mathematically derived, where we created two scenarios for each country: (1) only the reported cases imported from Wuhan were immediately isolated, but the rest of the estimated cases were not (main analysis) and (2) immediate isolation of 95% of the estimated imported cases from Wuhan. Here, we assumed that immediately isolated cases were not able to cause any secondary infection throughout their infectious periods, and hence the local outbreak risk estimation was conservative. For the main analysis, we truncated the uncertainty distribution of the imported case count derived earlier using the reported case count, to ensure that the estimated total number of cases imported from Wuhan was always greater than or equal to the reported case count. Countries with a local outbreak risk above 0.5 in our main analysis were named as potential donor countries, and subsequently assessed for their relative potential of exporting SARS-CoV-2 to any recipient country or territory outside mainland China, described as follows.