Effective containment explains sub-exponential growth in confirmed cases of recent COVID-19 outbreak in Mainland China

Benjamin F Maier; Dirk Brockmann

As we fixed the population size to be equal to the total population of the respective provinces, one might also ask how system size changes the results-for instance, if the outbreak is contained in a small region of a province, the effective population available to the transmission process will be substantially smaller. Therefore, we repeated the fit procedure with N as a free fit parameter to find that the form of X(t) is not altered substantially for different values of N, while I(t) can vary more strongly as reflected by a larger variation in I 0 . This result suggests that the estimation of the number of infecteds is associated with a larger uncertainty. The general shape of I(t), however, remains stable such that an inference of the peak time of unidentified infecteds is reasonable. Furthermore, the model favours larger values of the containment rate κ 0 for larger population sizes in order to reproduce emprical data. This is a reasonable because the number of available transmission pathways grows quadratically with the population size, making it easier for the disease to spread faster. In order to still observe sub-exponential growth, a large part of the susceptible population has to be removed quickly. Consistently, we find that a decay of susceptibles is necessary to reliably obtain the observed growth behavior (i.e. P > 0, κ 0 > 0).