Transmission interval estimates suggest pre-symptomatic spread of COVID-19

Lauren Tindale; Michelle Coombe; Jessica E Stockdale; Emma Garlock; Wing Yin Venus Lau; Manu Saraswat; Yen-Hsiang Brian Lee; Louxin Zhang; Dongxuan Chen; Jacco Wallinga; Caroline Colijn

Incubation period estimates were based on data for the earliest and latest possible exposure times, and on the reported times of symptom onset. It is impossible to confirm the exact times of exposure, so we use interval censoring (R package icenReg [16] ) to make parametric estimates of the incubation period distribution. Where the earliest exposure is not specified in the "start source" column, we noted that the case must have been exposed some time since the beginning of the outbreak (Dec 01, 2019). Where the last possible exposure is not given in "end source", we noted that exposure must have occurred before symptom onset. In the Singapore dataset, some cases had a travel history or contact with a known location or presumed source of the virus and this defined a most recent time of exposure. For these we proceed as in Tianjin. Other individuals had estimated exposure times based on the symptom times for their presumed infector. For these, we define an exposure window using the symptoms of their presumed infector ±4 days. Having defined exposure windows, we proceed with interval censoring as in the Tianjin data. In both datasets 5 . CC-BY-NC 4.0 International license It is made available under a author/funder, who has granted medRxiv a license to display the preprint in perpetuity.