Projecting the transmission dynamics of SARS-CoV-2 through the post-pandemic period

Stephen M Kissler; Christine Tedijanto; Edward Goldstein; Yonatan H. Grad; Marc Lipsitch

is the (which was not peer-reviewed) The copyright holder for this preprint . https://doi.org/10.1101/2020.03.04.20031112 doi: medRxiv preprint where b(t) is the strain-level incidence proxy on day t, and g(a) is the value of the generation interval distribution at time a. To translate the weekly incidence proxies to measures of daily incidence, we followed a previously described spline-based procedure (22) . For smoothing, the final weekly R is the geometric mean of the daily values in the preceding, current, and following weeks (moving 3-week average). We discarded the first three and last three weekly R estimates. To avoid unstable R estimates resulting from periods of low coronavirus activity, we The generation interval for the four commonly circulating coronaviruses has not been well-studied. In this analysis, we used the estimated serial interval distribution for SARS (Weibull with mean 8.4 days and standard deviation 3.8 days (32)) and varied this assumption in sensitivity analyses based on observations that the serial interval for currently circulating human coronaviruses and SARS-CoV-2 may be considerably lower (33) . We assumed that the maximum generation interval was the first day at which over 99% of the density had been captured.