Estimates of regional infectivity of COVID-19 in the United Kingdom following imposition of social distancing measures

Robert J Challen; Krasimira Tsaneva-Atanasova; Martin Pitt; Tom Edwards; Luke Gompels; Lucas Lacasa; Ellen Brooks-Pollock; Chris Martin; Gareth Griffith; Leon Danon

Essentially, R(t) is estimated by inferring connections between pairs of cases based on the difference in their dates of symptom onset and an assumed serial interval. The EpiEstim package implements the method outlined by Cori et al. (2013) 25 , which is based on an original method from Wallinga and Teunis (2004) 26 . The method uses a sliding time window during which the instantaneous reproductive number is assumed to be constant. We empirically decided on a 5 day sliding window for calculations of the time varying reproductive index, R(t), which aligns with the median incubation period of the virus, and offered a good trade off between noise and loss of detail. R(t) was visualised using ggplot and r-spatial libraries, using administrative maps from the UK Office for National Statistics 27 . The R(t) calculation was performed for the various regions of the NHS in England and compared to the overall time series for England, using a two sided t-Test to detect the significance of regional differences between the observations. . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.