An Investigation of Containment Measures Against the COVID-19 Pandemic in Mainland China
Ji Liu, Xiakai Wang, Haoyi Xiong, Jizhou Huang, Siyu Huang, Haozhe An, Dejing Dou, Haifeng Wang
AAn Investigation of Containment Measures Against theCOVID-19 Pandemic in Mainland China
Ji Liu , Xiakai Wang , Haoyi Xiong , Jizhou Huang , Siyu Huang , Haozhe An ,Dejing Dou , Haifeng Wang Baidu Inc., Beijing, 100085, China†These authors contributed equally to this work.* [email protected]
Abstract
As the recent COVID-19 outbreak rapidly expands all over the world, variouscontainment measures have been carried out to fight against the COVID-19pandemic. In Mainland China, the containment measures consist of three types,i.e., Wuhan travel ban, intra-city quarantine and isolation, and inter-city travelrestriction. In order to carry out the measures, local economy and informationacquisition play an important role. In this paper, we investigate the correlation oflocal economy and the information acquisition on the execution of containmentmeasures to fight against the COVID-19 pandemic in Mainland China. First, weuse a parsimonious model, i.e., SIR-X model, to estimate the parameters, whichrepresent the execution of intra-city quarantine and isolation in major cities ofMainland China. In order to understand the execution of intra-city quarantine andisolation, we analyze the correlation between the representative parametersincluding local economy, mobility, and information acquisition. To this end, wecollect the data of Gross Domestic Product (GDP), the inflows from Wuhan andoutflows, and the COVID-19 related search frequency from a widely-used Webmapping service, i.e., Baidu Maps, and Web search engine, i.e., Baidu SearchEngine, in Mainland China. Based on the analysis, we confirm the strongcorrelation between the local economy and the execution of information acquisitionin major cities of Mainland China. We further evidence that, although the citieswith high GDP per capita attracts bigger inflows from Wuhan, people are morelikely to conduct the quarantine measure and to reduce going out to other cities.Finally, the correlation analysis using search data shows that well-informedindividuals are likely to carry out containment measures.
Since December 2019, novel coronavirus COVID-19 has been identified and the outbreakhas expanded rapidly throughout tremendous countries, e.g., China [1], UnitedStates [2], European countries [3], etc. In China, the number of confirmed casesincreased from 571 on January 23, 2020 to 84,388 on May 1, 2020 and saturated around84.5 thousand. The COVID-19 has become a global emergency and is currentlyspreading throughout the whole world [4, 5]. In order to deal with the rapid outbreak ofthe COVID-19 pandemic in Mainland China, a range of containment measures havebeen put in place by Chinese authorities [6–8]. Similar containment measures have beenadopted in major countries all over the world [9, 10] etc.In Mainland China, the containment measures consist of intra-city and inter-citymeasures. As intra-city measures, suspected and confirmed cases have been quarantinedin hospitals or monitored self-quarantine at home [11], which is denoted the “quarantine”July 20, 2020 1/16 a r X i v : . [ phy s i c s . s o c - ph ] J u l - - - - - - - - - - . . . . . . N o r m a li z e d o u t f l o w s f r o m W uh a n - - - - - - - - - - - - Fig 1.
The comparisons of normalized outflows from Wuhan between 2020 and 2019 inMainland China. We believe the peak on April 4, 2019 is due to the vacation ofQingming Festival.measure in the paper. The authorities also encouraged citizens to stay-at-home anddiscouraged mass gatherings closed schools [12]. In addition, Wuhan city travel ban wasadopted, i.e., all transport was prohibited in and out of Wuhan city from 10:00 a.m. on23 January 2020, which incurred a significant reduction of the outflows from Wuhan asshown in 1. As shown in Figure 2, the national spring vacation has been prolonged andinter-city travel has been discouraged to reduce massive human migration across citiesin order to reduce infection. - - - - - - - - - - . . . . . . N o r m a li z e d v o l u m e o f m i g r a t i o n - - - - - - - - - - - - Fig 2.
The comparisons of normalized volumes of the volume of migration between2020 and 2019 in Mainland China. We believe the peaks between April 4, 2019 andApril 7, 2019 are due to the vacation of Qingming Festival.Mobile applications, e.g., Baidu Migration and search engines, e.g., Baidu , can beeasily used to achieve information acquisition for citizens to keep informed during theoutbreak of COVID-19. There are a great number of studies [1, 13–19] thatdemonstrated the feasibility to leverage mobile applications for information acquisition.As a result, the history search records can reflect the information acquisition status andthe history statistical migration data can be used to analyze the status of theCOVID-19 outbreak. As well-informed individuals are likely to travel less when there isCOVID-19 [20], it is interesting to analyze the correlation between informationacquisition and the execution of the containment measures. In addition, the correlation Baidu Migration - http://qianxi.baidu.com/ July 20, 2020 2/16etween the local economy and the information acquisition can be used to reveal howpeople in different economy situations react to the COVID-19 pandemic.In this work, we aim at using a parsimonious model, i.e., SIR-X model, and MarkovChain Monte Carlo (MCMC) [21] methods to estimate the parameters of the executionof intra-city containment measures in major cities of Mainland China. Then, we analyzethe correlation among different random variables, i.e., information acquisition status(COVID-19-related search frequency), local economy situation (GDP per capita) andthe parameters in the SIR-X model, in order to understand the relationship amongeconomy, information acquisition and the execution of containment measures. Morespecifically, we would like to investigate following problems: - - - - - - - - - - - - - - C o n f i r m e d C a s e s Official confirmed casesModel fit (a)
Number of confirmedcases in Beijing. - - - - - - - - - - - - - - C o n f i r m e d C a s e s Official confirmed casesModel fit (b)
Number of confirmedcases in Shanghai. - - - - - - - - - - - - - - C o n f i r m e d C a s e s Official confirmed casesModel fit (c)
Number of confirmedcases in Guangzhou. - - - - - - - - - - - - - - C o n f i r m e d C a s e s Official confirmed casesModel fit (d)
Number of confirmedcases in Shenzhen. - - - - - - - - - - - - - - C o n f i r m e d C a s e s Official confirmed casesModel fit (e)
Number of confirmedcases in Wuhan. - - - - - - - - - - - - - - C o n f i r m e d C a s e s Official confirmed casesModel fit (f)
Number of confirmed casesin Chongqing.
Fig 3.
Comparison between official number of confirmed cases and fitting number fromthe SIR-X model on May 1, 2020 in Beijing, Shanghai, Guangzhou, Shenzhen, Wuhanand Chongqing.•
How to construct a model to capture the confirmed cases?
This research issue hasbeen studied [7, 11] at province scale using the migration scale index released byBaidu Migration Open Data, where the index is calculated based on the pasthistorical statistical records from a widely-used Web mapping service, i.e., BaiduMaps . We propose to construct an SIR-X model based on the confirmed cases atcity scale using a Markov Chain Monte Carlo (MCMC) method. We provide solidresults using exact figures for major Chinese cities.• To what degree does the local economy affect the pandemic outbreaks of COVID-19and the execution of containment measures in major cities of China?
The impactof the COVID-19 pandemic on economy has been studied [22, 23] while thecorrelations between local economy and the outbreaks of the COVID-19 pandemicor the execution of containment measures are not analyzed. We hypothesized thatthe outflows from Wuhan trend to go to the cities where GDP per capital is high.We further hypothesized that there would be more initial confirmed cases as moreinfected people arrived at these cities. In addition, we hypothesized that thepeople in the cities where GDP per capital is high tend to perform moreinformation acquisition activities through voluntary COVID-19-related search inorder to be well informed on the situation of the COVID-19 pandemic. We analyzethe correlations based on the estimated parameters of SIR-X and the statisticaldata from Baidu Maps and provide solid results for major Chinese cities. Baidu Maps - https://map.baidu.com/
July 20, 2020 3/16
To what degree does the information acquisition affect the execution ofcontainment measures in major cities of China?
Strong positive correlationbetween the pandemic outbreaks of COVID-19 the information acquisition hasbeen reported in [20] while the correlation between the information acquisitionand the execution of containment measures are not analyzed. As reported in [20]and was seen in the collective responses to the emergencies [24] and panics [13],people voluntarily acquire information more frequently when the pandemicsituations become worse in their cities. We hypothesized the well-informed peopletend to apply the containment measures more strictly in order to avoid the risk tobe infected and the risk to make the situation worse. We analyze this correlationbased on the estimated parameters of SIR-X and the statistical data from BaiduMaps and Baidu Search Engine, and provide solid results for major Chinese citiesas well.Different from existing research [1, 6, 7, 11], we particularly analyze the correlationbetween local economy strength and the COVID-19-related search frequency with thecity population size (a controlling variable) removed in order to avoid the impact of thescale of city. Compared to the existing work [20, 22], we analyze the correlations notonly based on the data from Baidu Maps and Baidu Search Engine but also based onthe combination of an SIR-X model and MCMC methods.
In this section, we first present the existing models to capture COVID-19. Then, wepropose using SIR-X and MCMC to construct the model. Afterwards, we present thecomparison between the official number and the model fitting number of accumulatedconfirmed cases.Susceptible Infectious Recovered (SIR) model [25, 26] and Susceptible ExposedInfectious Recovered (SEIR) model [27–29] are largely adopted to characterize theoutbreak of COVID-19 epidemic. However, the containment measures cannot bedescribed in the standard SIR or SEIR model. A modified SEIR model [30] is proposedwith the consideration of mobility while it is still not able to infer the execution ofcontainment measures. A Long-Short-Term-Memory (LSTM) [30] model is proposed toproject the number of accumulated confirmed cases, which is not able to describe theexecution of containment measures either.In order to characterize the outbreak of COVID-19 epidemic with containmentmeasures at city level, we exploit the SIR-X model [11]. The SIR-X model is a modifiedSIR model, which takes the containment measures into consideration. We have the sameassumptions and use the same representative parameters as those in [11]. We assumethat there are public containment efforts, e.g., stay-at-home, reduced interaction withother people, which is referred as ‘containment’ and represented by a variable κ . Inaddition, we assume that infected individuals are quarantined, which is referred as‘quarantine’ and represented by a variable κ . We use α to represent the infection speedof an infected individual and β − to represent the average time an infected individualremains infectious before recovery or removal. Then, the SIR-X model is expressed bythe following differential equations: ∂ t S = − αSI − κ S∂ t I = αSI − βI − κ I − κI∂ t R = βI + κ S∂ t X = ( κ + κ ) I (1)July 20, 2020 4/16nstead of fixing the same parameters ( α and β ) for each province in [11], weestimated the parameters using a MCMC [21] method, inspired by [31]. In the model, I represents the number of initial infected individuals. The basic reproduction number R represents the average number of secondary infections an infected will cause beforehe or she recovers or is removed [11]. The reproduction number can be calculated as: R = αβ + κ + κ . We use R ,free to represent the reproduction number withoutcontainment or quarantine measures. As high temperature and high humiditysignificantly reduce the transmission of COVID-19 citeWang2020, R ,free and β may bedifferent for different cities because of the diversity of local environments. Thus, we usethe MCMC [21] method to estimate the distribution of the parameters, i.e., α , β , κ , κ , I , while the other parameters are fixed ( S is the population in the city, R (cid:48) is fixed as0 and X is the number of initial confirmed cases) at the beginning (January 23, 2020)with S , R (cid:48) , X and I representing the initial values of S , I , R and X . - - - - - - - - - - - - - - C o n f i r m e d C a s e s Official confirmed casesModel fit (a)
Aggregated number ofconfirmed cases in Hubei. - - - - - - - - - - - - - - C o n f i r m e d C a s e s Official confirmed casesModel fit (b)
Aggregated number ofconfirmed cases in Hebei. - - - - - - - - - - - - - - C o n f i r m e d C a s e s Official confirmed casesModel fit (c)
Aggregated number ofconfirmed cases inZhejiang. - - - - - - - - - - - - - - C o n f i r m e d C a s e s Official confirmed casesModel fit (d)
Aggregated number ofconfirmed cases in Sichuan. - - - - - - - - - - - - - - C o n f i r m e d C a s e s Official confirmed casesModel fit (e)
Aggregated number ofconfirmed cases inXingjiang. - - - - - - - - - - - - - - C o n f i r m e d C a s e s Official confirmed casesModel fit (f)
Aggregated number ofconfirmed cases in Yunnan.
Fig 4.
Comparison between official number of confirmed cases and fitting number fromthe SIR-X model on May 1, 2020 in Hubei, Hebei, Zhejiang, Sichuan, Xingjiang andYunnan. - - - - - - - - - - - - - - C o n f i r m e d C a s e s Official confirmed casesModel fit
Fig 5.
Comparison between official number of confirmed cases and fitting number fromthe SIR-X model on May 1, 2020 in China.Specifically, we use the uniform distribution as the parameter’s prior distribution.And with the consideration of the nonlinear of SIR-X model, we adopt the SequentialMonte Carlo sampler to achieve the posterior distribution of model’s parametersincluding the α and β . Finally, we take the expected value of each parameter toconstruct the model.July 20, 2020 5/16n order to have stable results from the MCMC method, we use a priori conditions,i.e., R < R (cid:48) ,free and κ < κ . The results of MCMC methods may not be stable, i.e.,the results of each execution may be different without a priori conditions. Thus, weintroduce a priori conditions, i.e., R < R (cid:48) ,free , κ < κ and the model fit number ofaccumulated confirmed cases should be equal or bigger than the official number ofconfirmed cases. R (cid:48) ,free represents a maximum value of R . We set R (cid:48) ,free as 6.2,which is in accordance with the result from [11] that the R should be between 1.4 and3.3. During the fitting process, if the a priori conditions are not met, the fitting processwill be executed again until reaching a limit, e.g., 20 times of execution, in order toavoid infinite execution. We assume that the quarantine measure is applied morestrictly on the infected individuals than other public citizens, i.e., κ < κ .In order to use the SIR-X model, we need to assume that few travelers andsymptomatic infected individuals travel into or from a city. As reported in [20], there isstrong correlation between the inflows from Wuhan and the confirmed cases in a city.We assume that few infected individuals travelled into major cities after January 23,2020 as Wuhan travel ban has been put in place since January 23, 2020 and few peoplewent to other cities from Wuhan as shown in Figure 1. In addition, we assume that thenumber of infected individuals in the inflows from other cities can be ignored comparedto the number of infected individuals among the local citizens in a city. With these twoassumptions, we can use the SIR-X and MCMC to estimate parameters for each majorcities in Mainland China based on the number of confirmed cases from January 23,2020 to May 1, 2020.Figures 3a - 3f illustrate the confirmed cases in several major cities of MainlandChina. From figures, we can see that the combination of SIR-X and MCMC captures thenumber of confirmed cases in different cities very well, e.g., Beijing, Shanghai, Shenzhen,Wuhan and Chongqing. However, the model does not well characterize the number ofconfirmed cases in Guangzhou as there are many (127 ) infected individuals from othercountries, which cannot be captured by the SIR-X model. In addition, we believe thatthe errors between the confirmed cases and fitted data are mainly due to the travelersfrom other countries in Beijing (174 confirmed cases from other countries) and Shanghai(326 confirmed cases from other countries). Then, we calculate the confirmed cases ofdifferent provinces by adding the number of confirmed cases in each affiliated city.Figures 4a - 4f shows the number of confirmed cases in several provinces of MainlandChina. We can see that the SIR-X well captures the aggregated cases at province scale.Furthermore, we use the same method to calculate the number of confirmed cases inMainland China as shown in Figure 5. Figure 6 shows that the combination of SIR-Xmodel and MCMC captures well the number of confirmed cases at city scale. Besides the number of confirmed cases (May 1, 2020), we collected three datasets, i.e.,GDP, mobility and COVID-19-related search frequency for major cities in MainlandChina. The GDP dataset was collected from [32], which characterized local economicdevelopment in 2019. The mobility data was captured from Baidu Maps and theCOVID-19-related search frequency (the ratio between COVID-19-related search volumefrom January to March 2020 and population in each city) data was gathered from awidely-used Web search engine. We analyze the correlation among local economy,mobility, search behaviors and the parameters estimated based on the combination ofthe SIR-X model and the MCMC method as presented in Section 2, for 238 cities in COVID-19 statistics - https://github.com/canghailan/Wuhan-2019-nCoV Confirmed cases in Guangzhou from National Health Commission - http://wjw.gz.gov.cn/ztzl/xxfyyqfk/yqtb/content/post_5815637.html
July 20, 2020 6/16 a b l e . O v e r a ll R e s u l t s o f C o rr e l a t i o n A n a l y s i s C o rr e l a t i o n s C o e ff . ( R ) p - v a l u e R e s u l t I G D P p e r c a p i t a v s . C O V I D - - r e l a t e d s e a r c h v o l u m e . % < . G D P p e r c a p i t a v s . C O V I D - - r e l a t e d s e a r c h f r e q u e n c y . % < . R e s u l t II G D P p e r c a p i t a v s . I n f l o w s f r o m W uh a n . % < . I n f l o w s f r o m W uh a n v s . I . % < . I v s . N u m b e r o f c o n f i r m e d c a s e s . % < . N u m b e r o f c o n f i r m e d c a s e s v s . C O V I D - - r e l a t e d s e a r c h f r e q u e n c y [ ] . % < . R e s u l t III G D P p e r c a p i t a v s . I n f l o w s f r o m W uh a n / p o pu l a t i o n . % < . I n f l o w s f r o m W uh a n / p o pu l a t i o n v s . R . % < . R v s . N u m b e r o f c o n f i r m e d c a s e s / p o pu l a t i o n . % < . N u m b e r o f c o n f i r m e d c a s e s / p o pu l a t i o n v s . C O V I D - - r e l a t e d s e a r c h f r e q u e n c y [ ] . % < . R e s u l t I V G D P p e r c a p i t a v s . κ . % < . C O V I D - - r e l a t e d s e a r c h f r e q u e n c yv s . κ . % < . R e s u l t V G D P p e r c a p i t a v s . O u t f l o w r ec o v e r y r a t e s − . % < . C O V I D - - r e l a t e d s e a r c h f r e q u e n c yv s . O u t f l o w r ec o v e r y r a t e s − . % < . July 20, 2020 7/16 a) The official number (May 1, 2020) ofconfirmed cases. (b)
The model fitting number (May 1, 2020)of confirmed cases.
Fig 6.
The comparison between official number and model fitting number of confirmedcases in major cities of Mainland China.
Fig 7.
Significant correlation among different factors.Mainland China (excluding Wuhan). We normalize inflow, outflow, search volume bythe following formula:
N ormalize (data) = data − data min data max − data min . (2)In this way, the data in the study are curved into the range from 0 to 1 proportionally.The results of our data-driven analysis are summarized in Table 1 and shown in Figure7. In this section, we present the observations obtained from the analysis. We have evidenced the significant positive correlations between local economy andCOVID-19-related search frequency for major Chinese cities (Result I in Table 1).
Inorder to analyze the correlation between two random variables, we calculated thePearson correlation coefficients [33] and conducted the Student’s T-test (two tails) toverify the significance test (the same for the following analysis in the paper). ThePearson correlation between the local GDP per capita and the total COVID-19-relatedsearch volume (between January and March 2020) is R ∗∗∗ = 52 . ( N = 238 and p -value = 3 . × − < . ) for each city. However, we considered that thisJuly 20, 2020 8/16 .0 0.2 0.4 0.6 0.8 1.0 GDP per capita S e a r c h f r e q u e n c y Fig 8.
Significant positive correlation has been found between the GDP per capita andCOVID-19-related search frequency for major cities in Mainland China. Shaded arearepresents the 95% Confidence Interval (CI).
GDP per capita I n f l o w r a t e Fig 9.
Significant negative correlations have been found between the GDP per capitaand inflow rate from Wuhan. Shaded area represents the 95% CI.observation can be incurred by the scale of the city, as a larger city would have largerpopulation, and would correspond to bigger GDP and high COVI-19-related searchvolume.We therefore tested the significance of the correlations between GDP per capita andCOVID-19-related search frequency, where we evidenced the significance in thecorrelations as R ∗∗∗ = 63 . ( N = 238 and p -value= . × − < . ). Inaddition, in order to obviate the impact of the scale of the city, i.e., the impact of citypopulation, we conducted partial correlation analysis [34, 35] between GDP per capitaand the COVID-19-related search frequency with the effects of the city population size(a controlling variable) removed. In order to estimate the partial correlation of random Inflows G D P p e r c a p i t a Fig 10.
Significant positive correlations have been found between the GDP per capitaand inflows from Wuhan. Shaded area represents the 95% CI.variables X and Y with the random variable Z removed, we expressed the partialcorrelation coefficient in terms of the Pearson correlation coefficients as ρ ( X, Y | Z ) ≡ ρ ( X, Y ) − ρ ( X, Z ) ρ ( Y, Z ) (cid:112) (1 − ρ ( X, Z ))(1 − ρ ( Y, Z )) . (3)July 20, 2020 9/16 .0 0.2 0.4 0.6 0.8 1.0 Inflows I Fig 11.
Significant positive correlations have been found between the inflows fromWuhan and I . Shaded area represents the 95% CI.We find a strong correlation with significance as well, such that R ∗∗∗ = 57 . ( N = 238 and p -value = 7 . × − < . ). Thus, we can conclude that no matter whetherthe scale of the city is big or small, the GDP per capita has a significant positivecorrelation with the COVID-19-related search frequency. Please see also in Figure 8 forthe visualization of the correlations. We have evidenced the significant positive correlation between the GDP per capita andinflows from Wuhan (Result II in Table 1).
We hypothesized that cities with higherGDP per capita would attract larger inflows from Wuhan. Therefore, for every city inthe study, we correlated the GDP per capita and the inflows from Wuhan, where weobtained Pearson correlation coefficients of R ∗∗∗ = 42 . ( N = 238 and p -value = 9 . × − < . ). In addition, in order to obviate the impact of the scaleof the city, we correlated GDP per capita and the inflows rate from Wuhan, i.e., theratio between inflows from Wuhan and the population. We found a strong positivecorrelation with significance as well, such that R ∗∗∗ = 32 . ( N = 238 and p -value = 2 . × − < . ). Please see also in Figures 9 and 10 for the visualizationof the correlations. Inflow rate R Fig 12.
Significant negative correlations have been found between the inflow rate fromWuhan and R . Shaded area represents the 95% CI. We have evidenced the significant positive correlation between the inflows and I (Result II in Table 1) and the significant positive correlation between the inflow rate and R (Result III in Table 1). We hypothesized that cities with larger inflows from Wuhanhave more initial infected cases, i.e., I in the SIR-X model. Thus, we correlated theinflows from Wuhan and I , where we obtained found a strong positive correlation withsignificance, such that R ∗∗ = 21 . ( N = 238 and p -value = 8 . × − < . ). Inaddition, we analyzed the correlation between the inflow rate and R , where we found astrong positive correlation with significance, such that R ∗ = 19 . ( N = 238 andJuly 20, 2020 10/16 .0 0.2 0.4 0.6 0.8 1.0 I0 N u m b e r o f c o n f i r m e d c a s e s Fig 13.
Significant positive correlations have been found between I and the number ofconfirmed cases. Shaded area represents the 95% CI. p -value = 2 . × − < . ). Please see also in Figures 11 and 12 for the visualizationof the correlations. We have evidenced the significance of the positive correlation between the number ofinitial infected individuals and the number of confirmed cases (Result II in Table 1) andthe positive correlation between R and the number of confirmed case rate (Result III inTable 1). We hypothesized that cities with bigger I finally have more confirmed cases.To this end, we performed correlation using I and the number of confirmed cases onMay 1, 2020. We found a significant positive correlation, such that R ∗∗ = 22 . with N = 238 and p -value = 4 . × − < . . In addition, we hypothesized that citieswith bigger R have more confirmed case rate, i.e., the ratio between the confirmedcases and the population. We performed correlation between R and the confirmed caserate, where we obtained a significant positive correlation, such that R ∗∗∗ = 29 . with N = 238 and p -value = 2 . × − < . . Please see also in Figures 13 and 14 forthe visualization of the correlations. R0 C o n f i r m e d c a s e r a t e Fig 14.
Significant negative correlations have been found between R and confirmedcase rate. Shaded area represents the 95% CI. G D P p e r c a p i t a Fig 15.
Significant positive correlations have been found between the GDP per capitaand κ . Shaded area represents the 95% CI.We obtained a strong positive correlation with significance between the number ofconfirmed cases and the COVID-19-related search frequency, such that R ∗∗∗ = 41 . July 20, 2020 11/16 N = 238 and p -value = 2 . × − < . ). Furthermore, we obtained a strongpositive correlation with significance between confirmed case rate and theCOVID-19-related search frequency, such that R ∗∗ = 21 . ( N = 238 and p -value = 9 . × − < . ). Similar results are also reported in [20].We thus can conclude that for each major city in the study, the GDP per capita andthe factors that incur infections, e.g., inflows from Wuhan, I , R , have significantpositive correlation. This indicates that the rich cities attract more inflows from Wuhan,which caused infections and in order to fight against COVID-19, the citizens in the richcities tend to perform more search activities in order to be well-informed. In this section, we analyze the correlation among local economy, information acquisitionand containment measures. As the quarantine measure (see details in Section 2) isdirectly related to the number of confirmed cases, we analyze the correlation between κ and other factors (local economy and information acquisition). In addition, we are alsointerested in the realization of inter-city containment measures, i.e., the outflow recoveryrate (the ratio between the outflows of 2020 and that of 2019). GDP per capita O u t f l o w r e c o v e r y r a t e s Fig 16.
Significant negative correlations have been found between the GDP per capitaand the outflow recovery rate. Shaded area represents the 95% CI. S e a r c h f r e q u e n c y Fig 17.
Significant positive correlations have been found between theCOVID-19-related search frequency and κ . Shaded area represents the 95% CI. We have evidenced the significance of the positive correlation between the executionof the quarantine measure and GDP per capita for major Chinese cities in the study(Result IV in Table 1).
Among all 238 cities in the correlation study, we hypothesizedthat people with high GDP per capita would try harder to realize the quarantinemeasure for infected individuals. Therefore, we correlated the GDP per capita and κ ,where Pearson correlation coefficients are R ∗ = 17 . ( N = 238 and p -value = 7 . × − < . ). Furthermore, we correlated the GDP per capita and theoutflow recovery rate, where we obtained Pearson correlation coefficients ofJuly 20, 2020 12/16 ∗∗∗ = − . ( N = 238 and p -value = 3 . × − < . ). The correlationanalysis result suggests that people with higher GDP per capita are more likely to applythe quarantine measure. Please see also in Figures 15 and 16 for the visualization of thecorrelations. We have evidenced the significance of the positive correlations between the realizationof quarantine measure and the COVID-19-related search frequency.
We performedcorrelation using the COVID-19-related search frequency and κ . We found a significantpositive correlation, such that R ∗ = 17 . with N = 238 and p -value = 5 . × − < . . Please see also in Figures 17 for the visualization of thecorrelations. We found negative correlations between the COVID-19-related searchfrequency and the outflow recovery rate are stronger with R ∗∗∗ = − . ( N = 238 and p -value = 1 . × − < . ) (similar result is also reported in [20]). Thecorrelation analysis result suggests that people with more per capita COVID-19-relatedsearch frequency are more likely to apply the containment measures, i.e., separate theinfected individuals and small outflow recovery rate.We can conclude that for every city in the study the GDP per capita and theCOVID-19-related search frequency have significant positive correlation to therealization of containment measures. We believe it is due to the will to avoid the risk tobe infected and the natural response to the fear and massive panics [13, 20]. In addition,the reason also goes to the fact that the people in the cities of higher GDP per capitatend to have bigger capacity or more tolerance to apply the containment measures. In this work, we first exploit the SIR-X model and MCMC method to estimate theparameters related to the COVID-19 pandemic at the scale of city. Then, we examinedthe correlation between the local economy and the spread of COVID-19 pandemic andthe execution of containment measures in major cities of Mainland China. Weconducted correlation analysis based on the mobility data and search data from BaiduMaps and Baidu Search Engine in Mainland China. Our analysis brings novelknowledge of the correlation among different factors related to the COVID-19 pandemic.The cities of higher GDP per capita attracts bigger inflows from Wuhan, which causemore confirmed cases. However, the demands of information from individuals becomeshigher, which incurs the reaction to apply the containment measures. Furthermore,well-informed individuals are more likely to apply the intra- and inter- city containmentmeasures, i.e., quarantine of infected individuals and reducing going to other cities. Theimplications of these correlations include that, the better the local economy is and themore that timely information acquisition is attained by residents, the better thecontainment measures are realized, which help fight against the COVID-19 pandemic inmajor cities of Mainland China.
Acknowledgement
To protect Baidu users’ data privacy, all experiments in this paper were carried outusing anonymous data and secure data analytics provided by Baidu Data FederationPlatform (Baidu FedCube). For data accesses and usages, please contact us via{fedcube, shubang}@baidu.com.July 20, 2020 13/16 uthor contributions statement
J. Liu formulated the research problems and drafted the manuscript. J. Liu, X. Wangand J. Huang collected data from Baidu, conducted the experiments, and performeddata analysis. S. Huang and H. An collected the consensus data and carried out thedata visualization. X. Wang, J. Huang, H. Xiong and D. Dou revised the whole paper.D. Dou and H. Wang proposed the research, coordinated the research efforts, andoversaw the whole research process.
References
1. Ahinazzi M, Davis JT, Ajelli M, Gioannini C, Litvinova M, Merler S, et al. Theeffect of travel restrictions on the spread of the 2019 novel coronavirus(COVID-19) outbreak. Science. 2020;368(6489):395–400.doi:10.1126/science.aba9757.2. University JH. Johns Hopkins University Coronavirus Resource Center; 2020. http://https://coronavirus.jhu.edu/ .3. Deslandes A, Berti V, Tandjaoui-Lambotte Y, Alloui C, Carbonnelle E, Zahar JR,et al. SARS-CoV-2 was already spreading in France in late December 2019; 2020.4. The Novel Coronavirus Pneumonia Emergency Response Epidemiology Team.The epidemiological characteristics of an outbreak of 2019 novel coronavirusdiseases (COVID-19) in China. China CDC Weekly. 2020;41(2):145.5. Sohrabi C, Alsafi Z, O’Neill N, Khan M, Kerwan A, Al-Jabir A, et al. WorldHealth Organization declares global emergency: A review of the 2019 novelcoronavirus (COVID-19). International Journal of Surgery. 2020;76:71–76.6. Kraemer MU, Yang CH, Gutierrez B, Wu CH, Klein B, Pigott DM, et al. Theeffect of human mobility and control measures on the COVID-19 epidemic inChina. Science. 2020;368(6490):493–497. doi:10.1126/science.abb4218.7. Tian H, Liu Y, Li Y, Wu CH, Chen B, Kraemer MU, et al. An investigation oftransmission control measures during the first 50 days of the COVID-19 epidemicin China; 2020.8. Ferretti L, Wymant C, Kendall M, Zhao L, Nurtay A, Abeler-Dörner L, et al.Quantifying SARS-COV-2 transmission suggests epidemic control with digitalcontact tracing; 2020.9. Engle S, Stromme J, Zhou A. Staying at Home: Mobility Effects of COVID-19;2020.10. Remuzzi A, Remuzzi G. COVID-19 and Italy: what next? The Lancet.2020;395(10231):1225 – 1228.11. Maier BF, Brockmann D. Effective containment explains subexponential growthin recent confirmed COVID-19 cases in China. Science. 2020;368(6492):742–746.doi:10.1126/science.abb4557.12. Chen S, Yang J, Yang W, Wang C, Bärnighausen T. COVID-19 control in Chinaduring mass population movements at New Year. Lancet.2020;395(10226):764–766. doi:10.1016/S0140-6736(20)30421-9.July 20, 2020 14/163. Ren SY, Gao RD, Chen YL. Fear can be more harmful than the severe acuterespiratory syndrome coronavirus 2 in controlling the corona virus disease 2019epidemic. World Journal of Clinical Cases. 2020;8(4):652.14. Yang Z, Zeng Z, Wang K, Wong SS, Liang W, Zanin M, et al. Modified SEIRand AI prediction of the epidemics trend of COVID-19 in China under publichealth interventions. Journal of Thoracic Disease. 2020;12(3):165.15. Li H, Chen X, Huang H. The novel coronavirus outbreak: what can be learnedfrom China in public reporting? Global Health Research and Policy.2020;5(1):1–3.16. Chen S, Yang J, Yang W, Wang C, Bärnighausen T. COVID-19 control in Chinaduring mass population movements at New Year. The Lancet.2020;395(10226):764–766.17. Zhang J, Lin G, Zeng J, Lin J, Tian J, Li G. Challenges of SARS-COV-2 andlessons learnt from SARS in Guangdong Province, China. Journal of ClinicalVirology. 2020;126:104341.18. Chen H, Chen Y, Sun B, Wang P, Wen L, Lian Z, et al. Correlation between themigration scale index and the number of new confirmed Novel CoronavirusPneumonia cases in China. Virus. 2020;11:3.19. Zhong P, Guo S, Chen T. Correlation between travellers departing from Wuhanbefore the Spring Festival and subsequent spread of COVID-19 to all provinces inChina. Journal of Travel Medicine. 2020;27(3):1–4.20. Xiong H, Liu J, Huang J, Huang S, An H, Kang Q, et al. Understanding theCollective Responses of Populations to the COVID-19 Pandemic in MainlandChina; 2020.21. Gilks WR. Markov Chain Monte Carlo; 2005.22. Huang J, Wang H, Xiong H, Fan M, Zhuo A, Li Y, et al.. Quantifying theEconomic Impact of COVID-19 in Mainland China Using Human Mobility Data;2020.23. Atkeson A. What Will Be the Economic Impact of COVID-19 in the US? RoughEstimates of Disease Scenarios. National Bureau of Economic Research; 2020.26867.24. Gao L, Song C, Gao Z, Barabási AL, Bagrow JP, Wang D. Quantifyinginformation flow during emergencies. Scientific reports. 2014;4:3997.25. Toda AA. Susceptible-Infected-Recovered (SIR) Dynamics of COVID-19 andEconomic Impact; 2020.26. Simha A, Prasad RV, Narayana S. A simple Stochastic SIR model for COVID 19Infection Dynamics for Karnataka: Learning from Europe; 2020.27. Peng L, Yang W, Zhang D, Zhuge C, Hong L. Epidemic analysis of COVID-19 inChina by dynamical modeling; 2020.28. Hou C, Chen J, Zhou Y, Hua L, Yuan J, He S, et al. The effectiveness ofquarantine of Wuhan city against the Corona Virus Disease 2019 (COVID-19): Awell-mixed SEIR model analysis; 2020.July 20, 2020 15/169. Tang S, Tang B, Bragazzi NL, Xia F, Li T, He S, et al. Analysis of COVID-19epidemic traced data and stochastic discrete transmission dynamic model; 2020.30. Yang Z, Zeng Z, Wang K, Wong SS, Liang W, Zanin M, et al. Modified SEIRand AI prediction of the epidemics trend of COVID-19 in China under publichealth interventions. Journal of Thoracic Disease. 2020;12(3):165–174.31. Wu JT, Leung K, Leung GM. Nowcasting and Forecasting the PotentialDomestic and International Spread of the 2019-nCoV Outbreak Originating inWuhan, China: A Modelling Study. Lancet. 2020;395(10225):689-697.doi:10.1016/S0140-6736(20)30260-9.32. of Statistics of China NB. National Bureau of Statistics of China; 2019.