Inferring COVID-19 spreading rates and potential change points for case number forecasts

Jonas Dehning; Johannes Zierenberg; Frank Paul Spitzner; Michael Wibral; Joao Pinheiro Neto; Michael Wilczek; Viola Priesemann

As median estimates, we obtain for the spreading rate λ = 0.41, µ = 0.12, D = 8.6, and I 0 = 19 (see Fig. 1D -H for the posterior distributions and the CIs). Converted to the basic reproduction number R 0 = λ/µ, we find a median R 0 = 3.4 (CI [2.4, 4.7]), which is consistent with previous reports that find median values between 2.3 and 3.3 [18, 33, 34] . Overall, the model shows good agreement for both new cases C t (Fig. 1 A) and the cumulative cases t t =0 C t (Fig. 1 B) with the expected exponential growth (linear in lin-log plot). The absolute deviation between data and model ( Fig. 1 C) is well captured by the case-number-dependent width of our likelihood (Methods) motivated by demographic noise in mean-field models of spreading processes [35, 36] . The observed data are clearly informative about λ, I 0 and σ (indicated by the difference between the priors (gray line) and posteriors (histograms) in Fig. 1 D,I,H) . However, µ and D are largely determined by our prior choice of parameters (histograms match gray line in Fig. 1 E,G) . This is to be expected for the initial phase of an epidemic outbreak, which is dominated by exponential growth.