Invisible spread of SARS-CoV-2 – Authors' reply
Adam J Kucharski
Rosalind M Eggo
Published:April 07, 2020DOI:https://doi.org/10.1016/S1473-3099(20)30275-9
We thank Nian Xiong and colleagues for their response to our Article.1 Although we separated individuals into exposed and infectious compartments in the basic model, we also considered a sensitivity analysis whereby people became infectious in the second half of their incubation period, and obtained the same conclusion (Article appendix p 12). We allowed the reproduction number, R, to vary over time in our model, rather than simply fix this value, to capture possible variation in transmission as a result of control measures and behaviour change. However, our median estimate for the reproduction number in mid-January of 2·4 is consistent with other estimates from the same period by use of a fixed R.2 As there is a delay from infection to symptom onset to hospitalisation, our model incorporated a delay to account for the time it takes for changes in transmission to be reflected in the observed data. Our estimate for transmission reduction was similar to that in another study, which focused on case counts in Wuhan and estimated that R had declined to around 1·3 by the last week of January, 2020.3 We disagree that our assumed incubation period was inappropriate; our assumption of a 5·2 day (SD 3·7) value is consistent with later studies that have estimated a similar value.4
