Transmission characteristics of the 2019 novel coronavirus
In order to reach between 1,000 and 9,700 infected cases by 18 January 2020, the early human-to-human transmission of 2019-nCoV was characterised by values of R0 around 2.2 (median value, with 90% high density interval: 1.4–3.8) (Figure 1). The observed data at this point are compatible with a large range of values for the dispersion parameter k (median: 0.54, 90% high density interval: 0.014–6.95). However, our simulations suggest that very low values of k are less likely. These estimates incorporate the uncertainty about the total epidemic size on 18 January 2020 and about the date and scale of the initial zoonotic event (Figure 2).
Figure 1  Values of R0 and k most compatible with the estimated size of the 2019 novel coronavirus epidemic in China, on 18 January 2020
The basic reproduction number R0 quantifies human-to-human transmission. The dispersion parameter k quantifies the risk of a superspreading event (lower values of k are linked to a higher probability of superspreading). Note that the probability density of k implies a log10 transformation.
Figure 2  Illustration of the simulation strategy, 2019 novel coronavirus outbreak, China, 2019–2020
The lines represent the cumulative incidence of 480 simulations with R0 = 1.8 and k = 1.13. The other parameters are left to vary according to the Table. Among these simulated epidemics, 54.3% led to a cumulative incidence between 1,000 and 9,700 on 18 January 2020 (in red).