Dynamic Estimation of Epidemiological Parameters of COVID-19 Outbreak and Effects of Interventions on Its Spread

Hongzhe Zhang; Xiaohang Zhao; Kexin Yin; Yiren Yan; Wei Qian; Bintong Chen; Xiao Fang

The last curve in the figure shows the estimated true number of actively infected but not quarantined cases by each day in the observation period (dotted black line). It refers to the number of actively infected people who are not quarantined at all (e.g., non-symptomatic infected cases 13 ) or not quarantined effectively (i.e., still being able to infect others). These infected people were not recorded by government reports either. Hence, we do not have the official number of actively infected but not quarantined cases. As shown, the estimated true number of actively infected but not quarantined cases peaked on February 7, 2020 (55,139 [95% CI 24, 273] ) and then started to decline. This decline was due to the operation of a number of new hospitals and a major COVID-19 testing facility 14 . As a result, many of those actively infected but not quarantined got tested and hospitalized. Fig. 2 plots the evolution of the effective reproduction number R in Wuhan from January 19, 2020 to February 24, 2020, with the shaded area representing the 95% credible interval. As discussed in the Method section, R is estimated using a rolling-window approach with 10-day window size. Therefore, R of day t indicates the transmissibility of COVID-19 in Wuhan over the time window of [t, t + 10]. Three major government measures illustrated in the figure include Wuhan lockdown effective January 23, 2020, which stopped all innercity and inter-city public transportations, vehicle ban effective January 26, 2020, which suspended all non-essential taxi, ride-hailing operation and private vehicle services, and large scale hospitalization beginning on February 5, 2020, which tested and hospitalized a large number of infected people due to added testing and treatment capacities. As shown in the figure, R of January 19, 2020 was 3.11 [95% CI 2.93-3.40]. It then climbed up and attained its maximum on January 24, 2020, which was 3.42 [95% CI 3.34-3.50]. This initial surge could be partly attributed to increased gathering and friend visiting during the period of the Chinese Spring Festival. The effective reproduction number R declined from January 24, 2020. This could be due to the two government measures that suspended transportation in Wuhan and subsequently reduced the average contact rate among Wuhan residents. The large scale hospitalization started on February 5 further reduced R and it became less than 1 from February 7, 2020 (0.76 [95% CI 0.65-0.92]).