Speaker
Description
We consider the problem of inferring $R_0$ for an epidemic where contact tracing data are available, from which we observe the size of several infectious clusters within a large population. By fitting the observed cluster sizes to the total progeny of a branching process, we use the Dwass formula for total progeny to estimate both $R_0$ and the index of dispersion of the offspring distribution. We extend this model to allow for the first generation of the branching process to have a different distribution to subsequent generations, mirroring the scenario in which traced contacts change their behaviour. We also explore right-censored cluster sizes and under ascertainment of cases, both of which can have appreciable impact on the resulting estimate of $R_0$. We provide results obtained by applying the method to FFX COVID-19 data from Albania which, in addition to synthetic data, are used to assess goodness of fit.
