key: cord-0915592-uqgrjjwp authors: Li, Xinhai; Zhao, Xumao; Sun, Yuehua title: The lockdown of Hubei Province causing different transmission dynamics of the novel coronavirus (2019-nCoV) in Wuhan and Beijing date: 2020-02-11 journal: nan DOI: 10.1101/2020.02.09.20021477 sha: dcd7a1235ea74e3ef71d051103bf8a64c3c8f457 doc_id: 915592 cord_uid: uqgrjjwp Background: After the outbreak of novel coronavirus (2019-nCoV) starting in late 2019, a number of researchers have reported the predicted the virus transmission dynamics. However, under the strict control policy the novel coronavirus does not spread naturally outside Hubei Province, and none of the prediction closes to the real situation. Methods and findings: We used the traditional SEIR model, fully estimated the effect of control measures, to predict the virus transmission in Wuhan, the capital city of Hubei Province, and Beijing. We forecast that the outbreak of 2019-nCoV would reach its peak around March 6±10 in Wuhan and March 20±16 in Beijing, respectively. The infectious population in Beijing would be much less (only 0.3%) than those in Wuhan at the peak of this transmission wave. The number of confirmed cases in cities inside Hubei Province grow exponentially, whereas those in cities outside the province increase linearly. Conclusions: The unprecedented province lockdown substantially suspends the national and global outbreak of 2019-nCoV. A novel coronavirus (2019-nCoV) appeared in December 2019 in Wuhan, Hubei 33 Province in central China had triggered city closure on Jan. 23, 2020, and lockdown 34 of all major cities in the province a few days later (Fig. 1 ). At present, over 50 million 35 people are constrained locally. Due to the threat of 2019-nCoV to public health, World 36 Health Organization (WHO) declared novel coronavirus (2019-nCoV) outbreak to be 37 "public health emergency of international concern" on Jan. 30, 2020 [1] 38 It is always a great challenge to fight effectively against a pandemic [2, 3] , 39 especially when little is known about the new virus [4] . Ideally, governments, 40 communities and medical services take rapid, effective, rational, and proportionate 41 responses to such health emergencies; either minimalist or maximalist responses may 42 potentially be very harmful [5, 6] . 43 To facilitate decision making against the 2019-nCoV, researchers had predicted 44 the transmission dynamics in different scenarios. Read et al. [7] estimated that only 45 5.1% (95% CI, 4.8-5.5) of infections in Wuhan were identified; ahead of 14 days, 46 they predicted the number of infected people in Wuhan to be greater than 250 47 thousand on Feb. 4, 2020. Read et al. [7] suggested, before the city closure of Wuhan, 48 that travel restrictions from and to Wuhan city are unlikely to be effective and 2019-49 nCoV would outbreak in Beijing, Shanghai, etc. with much larger sizes. Leung et al. 50 [8] estimated the transmission dynamics of 2019-nCoV in six major cities in China 51 under six scenarios: 0%, 25%, 50% transmission reduction with and without 50% 52 mobility reduction. However, the assumption of 50% mobility reduction is much 53 lower than the real situation. China government enforced tourism ban on Jan. 24, and 54 carried out other control measures such as extending holidays, closing schools, 55 cancelling meetings, suggesting a 14-day quarantine after travel. In particular, 56 highway traffic control is strictly implemented in many cities, towns, and villages. 57 The majority of communities in large cities such as Beijing and Shanghai have closed 58 to visitors. 59 Under the circumstance of strict control measures, we forecast the transmission 60 dynamics of Wuhan and Beijing. Wuhan is the source of 2019-nCoV, suffered a long 61 history (two months) of virus transmission, with lack of medical resources to 62 quarantine exposed and suspect people. Beijing is a larger city with 22 million 63 residents, including 10 million inbound passengers after the holidays of Chinese New 64 Year. We believe our estimation is close to the real situation and is helpful for 2019-65 nCoV control. 66 We used the Susceptible-Exposed-Infectious-Recovered (SEIR) model to estimate the 71 dynamics of the novel coronavirus (2019-nCoV). 72 The SEIR model has the form [9]: 74 . 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. For the cities with continuous imported infected people, we modify the growth 87 rate of infectious population as: 88 where imported is the daily number of imported infections. 90 The basic reproductive number R0 for 2019-nCoV had been estimated in several 92 independent studies (Table 1 ). It was noticed that changes in reporting rate of 93 confirmed cases substantially affected R0 estimation [11] . In fact, R0 is highly 94 . 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. The copyright holder for this preprint The mean value of R0 in Table 1 is 3.51, and the mean CV (coefficient of 98 variation) is 0.15. 99 The incubation period (1/α in SEIR model) was estimated as 6.4 (95% CI: 5.6 -100 7.7) days, ranging from 2. To make our analysis repeatable, we posted the data used in this study and R 112 code for the SEIR model at https://github.com/Xinhai-Li/2019-nCoV. 113 Until 9:00 on Feb. 9, the number of confirmed cases reaches 37251 115 (https://voice.baidu.com/act/newpneumonia/newpneumonia/?from=osari_pc_1). In 116 Hubei Province, the number of confirmed cases in the capital city Wuhan and other 117 . 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. The copyright holder for this preprint . https://doi.org/10.1101/2020.02.09.20021477 doi: medRxiv preprint prefecture-level cities grow exponentially; whereas in other cities in China, the 118 number of infections increase linearly (Fig. 2) . 119 . 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. The copyright holder for this preprint . https://doi.org/10.1101/2020.02.09.20021477 doi: medRxiv preprint We simulated the virus transmission process using the SEIR model and 124 estimated the model parameters for Wuhan (Fig. 3) and Beijing (Fig. 4) . For Wuhan, 125 we first fit the number of confirmed cases (Fig. 3A) , where the R0 is 5.75, much 126 higher than that in most studies yet lower than Tang et al.'s result (Table 1) imported patients from Jan. 19 to Feb. 8, which is 3.67. We adopted this uncertainty 147 to run 200 simulation, and found the R0 1.4 is the perfect fit (Fig. 4A) . Considering 148 the strict community locking (close to any visitors) and other extreme control 149 measures (such as office building control, i.e. checking body temperature for every 150 one and refusing visitors) in Beijing, we adopted a lower R0 1.2 after the first 30 days 151 of transmission. The uncertainty of R0 was simulated with CV = 15% (see Table 1 ), 152 and the results indicate that the infectious population may reach 607±553 on March 153 20±16 (Fig. 4B) . The uncertainty of the transmission dynamics, as simulated by the 154 SEIR model, is very high due to large variance in imported cases and R0 (Fig. 4) . 155 . 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. . 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. The copyright holder for this preprint . https://doi.org/10.1101/2020.02.09.20021477 doi: medRxiv preprint . 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. The copyright holder for this preprint . https://doi.org/10.1101/2020.02.09.20021477 doi: medRxiv preprint We simulated the transmission dynamics of 2019-nCoV, with taking into account of 174 the strict control measures enforced in the two cities. For Wuhan, city closure was a 175 chock upon the local residents and they took much better protection than before. 176 Accordingly, we halved the R0 after the city closure. As to Beijing, strict control 177 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. The copyright holder for this preprint . https://doi.org/10.1101/2020.02.09.20021477 doi: medRxiv preprint Feb. 6, 2020) had mentioned at his Wechat account that the nucleic acid detection for 205 him was negative, when he suffered breath difficulty. We think it would have a longer 206 fight (Fig. 3 & 4) Science. 2020. doi: doi:10.1126/science.abb1079. 234 . 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. (which was not peer-reviewed) The copyright holder for this preprint . https://doi.org/10.1101/2020.02.09.20021477 doi: medRxiv preprint Preparation for a pandemic: influenza A H1N1 Studies needed to address public health challenges of the 2009 238 H1N1 influenza pandemic: insights from modeling Influenza pandemic preparedness and severity 241 assessment of pandemic (H1N1) 2009 in South-east Asia H1N1, public health security, 245 bioethics, and human rights Influenza 248 pandemic preparedness in France: modelling the impact of interventions Novel coronavirus 252 2019-nCoV: early estimation of epidemiological parameters and epidemic 253 predictions. medRxiv Nowcasting and forecasting the Wuhan 2019-nCoV 255 outbreak. Preprint published by the School of Public Health of the University of 256 The mathematics of infectious diseases Distribution 260 and risk factors of 2009 pandemic influenza A (H1N1) in mainland china Preliminary estimation 264 of the basic reproduction number of novel coronavirus (2019-nCoV) in China, 265 from 2019 to 2020: A data-driven analysis in the early phase of the outbreak Transmission dynamics of 269 2019 novel coronavirus (2019-nCoV) Estimation of the 271 transmission risk of 2019-ncov and its implication for public health interventions. 272 Available at SSRN 3525558. 2020. 273 14. Backer JA, Klinkenberg D, Wallinga J. The incubation period of 2019-nCoV 274 infections among travellers from Wuhan Q&A on coronaviruses: How long is the incubation period? He X. The local infection of novel coronavirus in Beijing Beijing Business 278 Newspaper 280 Transmission of 2019-nCoV Infection from an Asymptomatic Contact in 281 Germany Study claiming new coronavirus can be transmitted by people 283 without symptoms was flawed