Summary
Background An ongoing outbreak of a novel coronavirus disease (COVID-19) was first reported in December 2019 in Wuhan city, Hubei Province, and has spread throughout China and to other countries. On 23 January 2020, in an attempt to contain the epidemic, non-essential travel was prohibited in and out of Wuhan city, a major transport hub and conurbation of 11 million people. Since then China has implemented nationwide its highest level (Level 1) of emergency response to further contain the spread of infection within and among cities.
Methods We used generalized linear regression models to investigate the effect of the type and timing of transmission control measures on the spread of COVID-19 from Wuhan city, and on the growth of the epidemic in 296 other cities across China. In addition to the Wuhan city shutdown, as part of the emergency response, entertainment venues were closed, public gatherings banned, intra-city public transport (bus and subway rail) suspended, and travel to and from other cities prohibited.
Findings The Wuhan city travel ban slowed the dispersal of infection to other cities by an estimated 2.91 days (95% CI: 2.54-3.29) on average. Among the other urban centres across mainland China, cities that implemented control measures pre-emptively, before the first case was reported, had 37% fewer cases in the week following the first reported case (13.0, 95%CI 7.1-18.8) compared with cities starting control after the first case (20.6, 95%CI: 14.5-26.8). Among individual control measures investigated, the most effective were suspending intra-city public transport, and closing entertainment venues and banning public gatherings.
Interpretation The implementation of transmission control measures slowed the dispersal of infection from its origin in Wuhan city and reduced the numbers of cases reported during the early stages of the epidemic in hundreds of other Chinese cities.
Evidence before the study We searched PubMed and preprint servers (medRxiv) using the terms “coronavirus”, “outbreak” and “intervention” regardless of language and date. 551 of the 571 results were unrelated to public health (“nonpharmaceutical interventions”). 13 studies of the remaining 20 studies used mathematical modelling to simulate the outbreak under various intervention scenarios; 5 were empirical analyses investigating the effect of interventions on disease in one or more cities, including the effect of traffic restrictions, school closure and influenza on the incidence of COVID-19. Two other papers evaluated the effect of transmission control measures in multiple cities on the 1918–19 influenza pandemic.
Added value of this study Our study is among the first study to evaluate the effects of an unprecedented, large-scale attempt to contain a rapidly-spreading, novel viral disease (COVID-19), including shutdown of the city of origin (Wuhan, 11 million inhabitants), followed by the closure of entertainment venues, the banning of public gatherings, the suspension of within-city public transport, and prohibition of travel to and from other cities across China. The results are derived from a large, newly compiled and geocoded repository of population and epidemiological data relevant to COVID-19.
Implications of all the available evidence Transmission control measures slowed the dispersal of COVID-19 from Wuhan city and reduced case numbers in hundreds of other cities as they became infected. All control measures appeared to reduce case numbers but the most effective were suspending public transport, closing entertainment venues and banning public gatherings. This analysis will help to inform the response to the epidemic in China and in other affected countries around the world.
Introduction
On 31 December 2019, less than a month before the Chinese New Year (Spring Festival), a cluster of pneumonia cases caused by an unknown pathogen were reported in Wuhan City, the largest transport hub in Central China. A novel coronavirus (COVID-19) 1, 2 was identified as the etiological agent 3, 4 and human-to-human transmission of COVID-19 has been since confirmed 5, 6. The increasing movement of people for the Chinese New Year was expected to spread the virus further in China and elsewhere, and by 13 February 2020 more than 50,000 COVID-19 cases had been reported from all provinces. To prevent further diffusion of COVID-19 from its source, Wuhan prohibited all transport in and out of the city from 10:00h on 23 January. To our knowledge, this is the largest attempted movement restriction or quarantine in human history with the aim of preventing infectious disease spread. By 23 January, China had raised its public health response level to its highest state of emergency (Level 1 of 4 levels of severity in the Chinese Emergency System, defined as an “extremely serious incident”).7 Here we present a quantitative analysis of the consequences and importance of some of the transmission control measures on the ongoing spread of COVID-19 across China.
As there is currently neither a vaccine nor a specific drug treatment for COVID-19, a range of public health (non-pharmaceutical) interventions has been used to control and mitigate the outbreak. To control transmission, suspected and confirmed cases have been isolated, public transport by bus and subway rail suspended, schools and entertainment venues have been closed, public gatherings banned, health checks carried out on migrants (“floating population”), travel prohibited in and out of cities, and information widely disseminated. Despite these measures, the epidemic has continued to spread geographically with mounting numbers of cases and deaths.
The measurement and evaluation of the effects of large-scale interventions are crucial to guide the response to this and future epidemics 8–13. To this end we have investigated, using the largest geocoded repository of population and city-level data relevant to COVID-19, the effect of control measures on the dispersal of infection from Wuhan city and on the number of cases reported from cities across China.
Methods
Data sources
Epidemiological and demographic data
We collected raw data from the daily official reports of the health commission of 34 provincial-level administrative units and 342 city-level units. Only laboratory-confirmed cases were used. We constructed a real-time database recording the date of the first reported case in all newly-infected cities with daily updates from 31 December 2019 to 6 February 2020. Population sizes for each city in 2018 were collected from the China City Statistical Yearbook (http://olap.epsnet.com.cn/). Using ArcGIS we calculated the great circle distance between Wuhan and each city reporting COVID-19 cases. The location of each city is geocoded by the latitude and longitude coordinates of the city centre.
Human mobility data
Human movement can be observed directly from mobile phone data, through the location-based services (LBS) employed by popular Tencent applications, such as WeChat and QQ. Movement outflows from Wuhan City to other cities, by air, train, and road, were calculated from the migration flows database (https://heat.qq.com/) 14 over the entire 2018.
Transmission control measures
After the Level 1 response was imposed, suspected and confirmed cases of COVID-19 were isolated and reported immediately in all cities. Given that available data for the city-level interventions in China are largely complete, we used the data for three transmission control measures : (i) closure of entertainment venues (169 cities with active interventions, 127 inactive), (ii) suspension of intra-city public transport (89 active, 207 inactive), and (iii) prohibition of travel by any means to and from other cities (171 active, 125 inactive). Cities that had reported cases for less than 7 days were not included, a total of 45 cities (plus Wuhan) were excluded. Each city was regarded as implementing an intervention when the official policy was announced to the public. Other transmission control measures, such as delineating control areas, closure of schools, isolation of suspected and confirmed cases, disclosure of information, were not investigated in our analysis because they were reported to have been implemented in all cities uniformly and without delay. The timing of different control measures implemented in each city was recorded, including the implementation time delay since 31 December 2019; a shorter time delay corresponds to earlier implementation.
Data analysis
Effect of the Wuhan city shutdown on the spread of infection to other cities
In order to quantify the effect of the Wuhan travel shutdown (23 January 2020) on COVID-19 epidemic spread, we used data collected between 31 December 2019 and 28 January 2020. The association between distance, human movement, interventions and epidemic timing of COVID-19 was assessed with a regression analysis using a General Linear Model (GLM). Among five possible regression models examined, the best model (judged by the Akaike Information Criterion) was: where TotalFlowj represents the passenger volume from Wuhan City to city j by airplane, train, and road. Popj is the population in city j. Latj and Lon j represent latitude and longitude of city j. The dummy variable shutdownj is used to identify whether the arrival time of the newly infected city j is influenced by the shutdown of Wuhan, where 0 represents a “no intervention scenario” and 1 represents intervention. The dependent variable Yj is the arrival time of the epidemic in city j, which measures the spatial spread of COVID-19. The βi are the regression coefficients.
Effect of transmission control measures adopted by other cities
In this early evaluation, carried out as the epidemic unfolds, the effect of control was measured in terms of the total number of cases confirmed during the first week of the epidemic in each city, starting from the day on which the first case was reported. Mann-Whitney U tests were used to compare the total numbers of cases in cities that implemented control measures before and after the first case was reported.
The association between the type and timing of transmission control measure (TCM) and epidemic intensity was evaluated with a Poisson regression model. Controlling for city population size and the influx of travellers from Wuhan: where i is a given city and j refers to each TCM. The dependent variable μ is the total number of cases reported during the first week of the epidemic in each city. The βi and βji are regression coefficients. The binary variable TCMj is used to identify whether the jth transmission control measure was implemented by the corresponding city (0 represents no-intervention and 1 represents intervention). TTCMj represents the timing of the jth control measure. TCMj × TTCMj models the interaction between the timing and implementation of each TCM. Dis is the great-circle distance from city i to Wuhan City. ArrT is the arrival time of the epidemic in city i, which measures the spatial spread of COVID-19. The regression analysis was performed using the MASS routine in the R package (R version 3.4.0).
Results
Between 31 December 2019 and 6 February 2020, interventions were adopted by 342 cities in China, including Wuhan (Figure 1). In order to quantify the effect of the Wuhan travel shutdown on epidemic spread we analysed the arrival time of COVID-19 from Wuhan to each city as a function of geographic distance (between city centres) and of human movement by air, train, and road, as recorded by Tencent’s location-based services database. The dispersal of COVID-19 was rapid (Figure 2A): 262 cities reported cases within 28 days. For comparison, the influenza H1N1pdm pandemic in 2009 took 132 days to reach the same number of cities in China (Figure 2A). Most cities with early arrival dates were in southeast China among which there is greater mobility and higher population density. The number of cities providing first reports of COVID-19 peaked at 59 on 23 January (the day of the Wuhan travel ban).
We quantified the degree to which the Wuhan city travel ban slowed disease spread (Table 1). COVID-19 arrived sooner in those cities that had larger populations and had more travellers from Wuhan. On average, the Wuhan city shutdown delayed the arrival time of COVID-19 in other cities by 2.91 days (95%CI 2.54-3.29 days). Without the shutdown, the expected arrival time of COVID-19 from Wuhan would have been 22.3 days (95%CI 20.6-24.0 days) from 31 December 2019 (Figure 2B). More than 130 cities, covering more than half the geographic area and population of China, benefited from this intervention (Fig. 2C).
Table 2 shows the timing and implementation of transmission control measures in the 342 cities. Each city adopted at least one category of control measure and 136 cities applied all three measures (Figure 3). School closure combined with the isolation of suspected and confirmed patients and with the disclosure of information was implemented in all 342 cities. Public gatherings were banned and entertainment venues closed in 220 cities (64.3%). Intra-city public transport was suspended in 136 cities (39.7%) and inter-city travel prohibited by 219 cities (64.0%).
Cities that implemented a Level 1 response (any combination of control measures) before the first case was confirmed had 37% fewer cases in the week following that first case (13.0 cases, 95%CI: 7.1-18.8, n#125) compared with cities that started control after the first case was confirmed (20.6 cases, 95%CI: 14.5-26.8, n#171) (U#8197 z#-3.4, P<0.01; Figures 4 and 5). Cities that suspended intra-city public transport and/or closed entertainment venues and banned public gatherings, and did so sooner, reported, fewer cases during the first week (Table 3). There is no evidence that the prohibition of travel between cities, which was implemented after the Wuhan shutdown on 23 January, reduced the number of cases in other cities across China (Table 3).
Discussion
This is the first study, to our knowledge, to evaluate the effects of an unprecedented and large-scale attempt to contain the spread of COVID-19. Interventions included shutdown of the city of origin (Wuhan), followed by the closure of entertainment venues, the suspension of within-city public transport, and prohibition of travel to and from 341 other cities across China.
These results are derived from a large geocoded repository of population and epidemiological data relevant to COVID-19. They suggest that the combination of transmission control measures slowed the dispersal of COVID-19 from Wuhan city and reduced the numbers of cases reported in hundreds of other Chinese cities in the first week after those cities had recorded their first COVID-19 cases. All control measures had significant effects within cities but the most effective, according to this analysis, were suspending public transport and closing entertainment venues. These control measures may not have reduced the overall size of the epidemic, but they did give some extra time to reinforce the response.
At least three caveats must be attached to this early evaluation of control measures. First, we cannot prove that the control measures investigated here caused the reported effects because these interventions may have stimulated other unrecorded behavioural changes that mitigated the spread of infection. Second, we could not investigate the effects of control measures that were said to have been implemented uniformly and without delay in all cities. Among these were: the identification of affected areas in cities, school closures, and the isolation of suspects and other patients with infectious diseases. Third, the interventions that we could investigate explained only part of the variation in case numbers reported among Chinese cities. Further data and analysis are therefore needed in order to distinguish the impact of individual control measures, and to understand the causes of the remaining variation.
Urbanization and the development of rapid transport systems in China15–18 probably accelerated the spread of COVID-19 across China, as suggested by the comparatively slow dispersal of pandemic influenza H1N1pdm in 2009. The control measures reported here have significantly slowed the COVID-19 epidemic, but additional measures are clearly needed, including more rapid detection and isolation of patients with any signs or symptoms (rather than waiting for a confirmed diagnosis), more intensive contact tracing, and the rapid detection of cases as they appear in new locations in China.
Data Availability
Population sizes for each city were collected from the China City Statistical Yearbook (http://olap.epsnet.com.cn/).Human movement can be observed directly from mobile phone data, through the location-based services (LBS) employed by popular Tencent applications, such as WeChat and QQ. Average movement outflows from Wuhan City to other cities, by air, train, and road, were calculatedfrom the migration flows database (https://heat.qq.com/) over the entire 2018.
Acknowledgements
We thank the many thousands of CDC staff and local health workers in China who collected data and continue to strive to contain COVID-19 in China and elsewhere. Funding for this study was provided by the Beijing Natural Science Foundation (JQ18025); Beijing Advanced Innovation Program for Land Surface Science; National Natural Science Foundation of China (81673234); Young Elite Scientist Sponsorship Program by CAST (YESS)(2018QNRC001); HT, OGP and CD acknowledge support from the Oxford Martin School; H.T. acknowledges support from the Military Logistics Research Program. The funders had no role in study design, data collection and analysis, the decision to publish, or in preparation of the manuscript.
Footnotes
Author contributions: H.T., P.Z., R.F.Y., O.G.P., B.G., C.D. designed the study. B.C. and Y.M.S. collected and processed the Tencent’s LBS data. Y.H.L., B.Y.L., B.X., Q.Q.Y., B.W., P.Y., Y.J.C., Q.Y.W. collected the statistical data. H.Y.T., Y.L., C.H.W, and J.C. conducted the analyses. M.K., O.B., R.F.Y., O.G.P., B.G., and C.D. edited the manuscript. H.T. and Y.D.L. wrote the manuscript. All authors read and approved the manuscript.