key: cord-0737817-y23uvvut authors: Yu, Shanshan; Cui, Shufeng; Rui, Jia; Zhao, Zeyu; Deng, Bin; Liu, Chan; Li, Kangguo; Wang, Yao; Yang, Zimei; Li, Qun; Chen, Tianmu; Wang, Shan title: Epidemiological Characteristics and Transmissibility for SARS-CoV-2 of Population Level and Cluster Level in a Chinese City date: 2022-01-18 journal: Front Public Health DOI: 10.3389/fpubh.2021.799536 sha: ebe2d4be0af478f6f1f7230011ec3ae44aeb790b doc_id: 737817 cord_uid: y23uvvut BACKGROUND: To date, there is a lack of sufficient evidence on the type of clusters in which severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) is most likely to spread. Notably, the differences between cluster-level and population-level outbreaks in epidemiological characteristics and transmissibility remain unclear. Identifying the characteristics of these two levels, including epidemiology and transmission dynamics, allows us to develop better surveillance and control strategies following the current removal of suppression measures in China. METHODS: We described the epidemiological characteristics of SARS-CoV-2 and calculated its transmissibility by taking a Chinese city as an example. We used descriptive analysis to characterize epidemiological features for coronavirus disease 2019 (COVID-19) incidence database from 1 Jan 2020 to 2 March 2020 in Chaoyang District, Beijing City, China. The susceptible-exposed-infected-asymptomatic-recovered (SEIAR) model was fitted with the dataset, and the effective reproduction number (R(eff)) was calculated as the transmissibility of a single population. Also, the basic reproduction number (R(0)) was calculated by definition for three clusters, such as household, factory and community, as the transmissibility of subgroups. RESULTS: The epidemic curve in Chaoyang District was divided into three stages. We included nine clusters (subgroups), which comprised of seven household-level and one factory-level and one community-level cluster, with sizes ranging from 2 to 17 cases. For the nine clusters, the median incubation period was 17.0 days [Interquartile range (IQR): 8.4–24.0 days (d)], and the average interval between date of onset (report date) and diagnosis date was 1.9 d (IQR: 1.7 to 6.4 d). At the population level, the transmissibility of the virus was high in the early stage of the epidemic (R(eff) = 4.81). The transmissibility was higher in factory-level clusters (R(0) = 16) than in community-level clusters (R(0) = 3), and household-level clusters (R(0) = 1). CONCLUSIONS: In Chaoyang District, the epidemiological features of SARS-CoV-2 showed multi-stage pattern. Many clusters were reported to occur indoors, mostly from households and factories, and few from the community. The risk of transmission varies by setting, with indoor settings being more severe than outdoor settings. Reported household clusters were the predominant type, but the population size of the different types of clusters limited transmission. The transmissibility of SARS-CoV-2 was different between a single population and its subgroups, with cluster-level transmissibility higher than population-level transmissibility. The global pandemic of the coronavirus disease 2019 (COVID- 19) has had a serious impact on public health systems. COVID-19 differs significantly from infectious diseases such as SARS and influenza in terms of epidemiological characteristics (1) . As for the transmissibility, COVID-19 is more infectious than influenza, but weaker than measles (2) . Of those people who presented with symptoms significant enough to be classed as patients, the majority (81%) presented with mild to moderate symptoms, while 14% presented with severe symptoms, and 5% suffer critical symptoms (3) . Older people are at a higher risk of developing severe symptoms (4) . It is necessary to estimate the transmissibility of COVID-19 to determine the severity and size of the pandemic, and to design appropriate interventions and responses to protect the population and control the spread of the disease (5) . A systematic review found that the population-level transmissibility of SARS-CoV-2 was 3.32 (95% CI, 2.81 to 3.82) (6) , with WHO estimates ranging from 1.4 to 2.5 (7) . The transmissibility of the Delta variant, is substantially higher. Among five studies cataloged in October 2021, the mean estimated basic reproduction number (R 0 ) was 5.08 for Delta (8) . Most studies for cluster-level epidemic focused on the epidemiological characteristics of transmission chains and case profiles across generations (9) (10) (11) . Researchers have found many examples of SARS-CoV-2 clusters associated with indoor settings, with many reports from households, few from schools, and increasing reports from hospitals and elderly care settings across Europe (12) . Limited studies have focused on the transmissibility of the virus in cluster-level outbreaks. A SEIR (susceptibleexposed-infectious-removed) based modeling study calculated an R 0 of 3.06 (95% CI: 2.64-3.51) for 15 clusters (13) . Another cluster-based study found a mean transmission rate of 1.86 per case among family members (14) . To amplify the reasons for the reported heterogeneity in transmission: the number of people infected by one infected person generally varies (15) ; as only 10 to 20% of the population are responsible for the disease's spread (16) . It often spreads in clusters, where infections can be traced back to an index case or geographical location (17) . In these instances, superspreading events often occur, where many people are infected by one person (15) . Thus, we need to know in which types of clusters the virus is most likely to be transmitted. Notably, the differences in the epidemiological characteristics and transmissibility of cluster-level and population-level outbreaks remain unclear. Following the current removal of suppression measures in China, identifying the characteristics of both levels including epidemiology and transmission dynamics allows us to develop better surveillance and control strategies. In this study, we first compared and analyzed the differences in epidemiological characteristics of COVID-19, including the population distribution, incubation period, and time interval. We then proposed the SEIAR (Susceptible-Exposed-Infectious-Asymptomatic-Removed) model (18) (19) (20) for calculation of population-level transmissibility (PLT). After a rigorous investigation to clarify the relationship between primary and secondary cases, we calculated the cluster-level transmissibility (CLT) by definition (21) . CLT is defined as the expected number of cases infected by a single case during the entire infectious period. This study was divided into four sections (Figure 1 ). The first section briefly described the epidemiological characteristics of reported COVID-19 cases, including temporal, geographical, age and gender distributions, source of infection, clinical severity as well as incubation period and time interval. In the second section, the SEIAR model was established for calculating R eff as PLT (i.e., population-level transmissibility). The calculation of CLT (i.e., cluster-level transmissibility) by definition of R 0 were presented in the third section. The final section compared cluster-level and population-level transmissibility. In this study, we collected daily reported COVID-19 cases in Chaoyang District (Figure 2 ) from 1 January 1 2020 to 31 December 2020 from the Chinese Center for Disease Control and Prevention through the National Notifiable Disease Surveillance System (22) . The variables included in the COVID-19 dataset mainly comprised gender, age, occupation, the site of residence, symptomatic infections, asymptomatic infections, date of onset, date of diagnosis, date of discharge or death, and severity of disease. Demographic data were obtained from the Chinese Statistical Yearbook. For definitions of symptomatic and asymptomatic infected cases and clinical types, we refer to the Prevention and Control Program for COVID-19 (23) published by the National Health Council. In our study, types of clusters were defined according to where they occurred ( Table 1 ). There are basically three types of places where a person stays: one leaves his or her home and go to work, and all the places he or she passes through are called communities. Clusters at the household level imply that these outbreaks occurred at home. Factory-level clusters refer Household Transmission between individuals in a shared living space In the workplace, typically an office. Where transmission occurs on public property and does not fall into any of the above two settings e.g., supermarkets, buses, hotel, park, etc. to places like offices or schools, which consist of various places where people work. Community-level clusters mean that these outbreaks occur anywhere outside the home and workplace, such as supermarkets, buses, etc. We take the actual data from 1 January 1 2020 to 2 March 2020 to fit the SEIAR model. Our model (Figure 3 ) was built based on the approach described in our previous paper (18) (See Supplementary Materials for detailed information of this model). The estimation of the parameters and the initial values of the variables were shown in Table 2 . We calculated the PLT using this SEIAR model. At population-level epidemics, the population was not completely susceptible or the intervention measures were taken, the effective reproductive number (R eff or R t ) was used to express the transmissibility of infectious diseases. R eff is defined as the expected number of secondary cases arising from a single infected individual at time t, with a given level of immunity in the population. In this study, R eff is calculated by the secondgeneration matrix method: where β denotes the infection rate coefficient, which is the average number of contacts per person per time multiplied by the probability of disease transmission in a contact between susceptible and an infectious subjects; S is the susceptible population; p denotes the proportion of the asymptomatic infections; γ denotes the probability of an infectious individual recovering within the average infectious time period, and 1/γ means the infectious period of symptomatic infections; f means the fatality of the disease; κ indicates the relative transmission rate of asymptomatic infections to symptomatic infections. At cluster-level epidemics, the basic reproduction number (R 0 ) was suitable for evaluating the transmissibility quantitatively. Based on the nine clusters according to the inclusion and exclusion criteria (See Supplementary Figure S1 for detailed information), we aimed to calculate the CLT based on the definition of R 0 (See Supplementary Materials for definition of R 0 ). This calculation method required a detailed investigation to clarify the relationship between the cases which was available in studies of clusters. We directly divided the number of cases in the (n + 1) generation by the number of cases in the (n) generation. For example, Ebola has an R 0 of two, so on average each person with Ebola passes it on to the two other people (See Supplementary Figure S2 ). The formula is as follows, (2) where N denotes the number of cases, N (n+1)th means the number of cases in the (n+1) generation, N nth means the number of cases in the (n) generation. Cases mean symptomatic or asymptomatic infections. Microsoft In 2020, the COVID-19 outbreak in Chaoyang District was divided into three different stages of prevention and control (Figure 4 ). The first stage (January 1 2020 to February 29 2020) was the control of cases from Chinese cities including Wuhan City. The second stage (March 1 2020 to May 31 2020) was to control cases from abroad. The third stage (June 1 2020 to December 31 2020) was to control local clusters caused by imported cases. We found a balanced gender distribution (sex ratio = 104:107) among COVID-19 cases ( Table 3) . The majority of patients were between 20 and 59 years of age, with a median age of 33 years. Critical and severe cases (21/211, 10%) were rare. The rate of asymptomatic infection was 0.004 per 1,000 people. For the epidemiological parameters (Table 4) , we set a fixed incubation period of 5.0 to 7.0 days [d] for symptomatic and asymptomatic cases of infection, respectively, based on previous study (23) . Four clusters (44.4%) reported only one secondary case and another five clusters (55.6%) reported more than two secondary cases. The size of each cluster ranged from 2 to 17 cases (M = 2). In 2020, clusters occurred from the initial household to the factory to the community (Figure 5) . Table 5) . The modeling results showed (Figure 6 ) that R eff reached 4.81 before February 1, that is, an infected person can infect an average of 4.81 susceptible persons during the average incubation period. After February 1, R eff decreased to 0.81, with a reduction of almost 100%, indicating that the outbreak was gradually controlled. As for the household-level cluster, the median R 0 was one (Figure 7) . Sixteen individuals were diagnosed with COVID-19 in this factory-level cluster, with R 0 of 16 (Figure 8) . Three individuals in this community-level cluster were diagnosed with COVID-19, with R 0 of three (Figure 9) . The factory-level cluster (R 0 = 16) had higher transmissibility than the community-level cluster (R 0 = 3) and the household-level cluster (R 0 = 1). We found that the population-level epidemic was characterized by three stages with different primary prevention and control populations, which was similar to previous studies (26) . In our study, we set 5-7 d (23) as the population-level incubation period, which was consistent with previous work. The results of Lauer The setting where clusters occurred also changed, from household to factory and to community. In the early stage, cold weather and isolation measures contributed to the survival of the virus and the increased risk of intra-household transmission (31, 32) . The long chain of transmission at factorylevel cluster could be the failure to effectively disinfect and maintain social distances once people started returning to the workplace. Outbreaks occur in different workplaces when it is difficult to maintain the recommended distance of at least 2 meters (33, 34) . Shared facilities (e.g., canteen and dressing rooms), transportation and accommodations may also lead to transmission (35) . Coronavirus infections are characterized by intermittent shedding, long incubation periods, and sampling locations can affect detection results, therefore there were cases undetected (36), causing a community-level transmission after removal of isolation at the end of 2020. For household clusters, the length of the incubation period was similar to previous studies (4, 37) . The longer incubation period for this factory-level cluster may be related to the increase in asymptomatic infections, individual heterogeneity, and virus mutation. This community-level cluster occurred at a late stage. At this time, the intensity and efficiency of nucleic acid detection has improved considerably compared to the previous stage, so that the time to detect cases is shorter than that in the earlier household-level clusters. In addition, the discrepancy may be due to lack of accuracy in the investigation of the time of contact with the infectious agent and lack of large sample size. Our R eff value was higher than those estimated by the SAPHIRE (Susceptible-unascertained-presymptomatic-hospitalizedinfectiousness-infectious-recovered-exposed) model for early outbreaks in Wuhan City (38) , and by World Health Organization (R eff = 1.4-2.5) and by other studies for Beijing City (R eff < 3) (7, 39, 40). While our estimation was similar to the values estimated for the SARS (severe acute respiratory syndrome) epidemic in Beijing, China (R 0 = 4.91) (41) , and for MERS (Middle East respiratory syndrome) in Jeddah, Saudi Arabia (R 0 = 3.5 to 6.7) (42). There are several possible reasons for this. First, this high R eff was calculated for the early stages, where preventive measures were inadequate due to the small sample size of cases and the lack of awareness about the disease. Additionally, the estimates depend on social and demographic variables, the estimation method used, the validity of the underlying assumptions, and the biology of the infectious agent. Third, modeling methods also contribute to the difference (43) . Finally, estimates of R eff may be wrong due to insufficient data, thus precisely estimating R eff is rather difficult. The setting with the highset number of reported clusters of SARS-CoV-2 transmission was households, but householdlevel transmissibility (R 0 = 1) was not the highest. The basic reproduction number may be limited by the number of people in each cluster. For example, the number of family members determines the upper limit of secondary cases in household transmission. Several outbreak investigation reports suggest that SARS-CoV-2 transmission can be particularly effective in crowded, confined indoor spaces (12) . Patients involved at the factory-level were more effective in contacting susceptible populations than at the household-level, leading to high transmissibility. Multiple outbreaks of COVID-19 have been observed in several occupational settings, including slaughterhouses, meat processing plants, mines and building sites (12, 44) . Possible factors contributing to clusters in occupational settings are listed below. Studies have shown that in Europe, more than 80% of working time is spent indoors, and changes in socioeconomic and demographic conditions have led to different work-day patterns indoors (45) . Participating in meetings and sharing the same office space has been reported in literature as a risk factor for contracting COVID-19 (33, 46) . Both population-level and cluster-level epidemic were characterized by multiple stages. The incubation period and time interval were similar to previous studies. Populationlevel transmissibility (R eff = 4.81) was intermediate to cluster level (R 0 = 1∼16). Frequent close contacts more or less create opportunities for increased transmission risk, such as indirect contact transmission and respiratory droplet transmission. R 0 values differed in different subgroups of a single population. In fact, the total value of R eff in a population is the average of the R 0 subtypes in that population (47) . We need to note that even if the total value of R 0 in a population is moderate (R eff = 4.81), the transmission potential of some subgroups in that population may still be high, for instance, the factory-level R 0 is equal to 16. Besides, the basic reproduction number is affected by several factors, including the duration of infection of the affected person, the infectiousness of the microorganism, and the number of susceptible individuals in the population with which the infected people is in contact (48) . In the early stage of the disease outbreak, R eff was estimated based on a differential equation model of the overall epidemic data. However, it only represents the average epidemic pattern, and it ignores the heterogeneity of infectivity within populations due to structural differences in different settings such as households, factories and communities. Based on detailed data obtained from epidemiological investigation, R 0 can be calculated by definition. This indicator excludes the possibility of close contacts contacting other infected persons and therefore has some advantages in analyzing the transmissibility of clusters. Following the current lifting of suppression measures in China, the key population is mainly imported cases, and asymptomatic infections. Active screening and expanded nucleic acid testing for the population and high-risk infection subgroups are needed, especially indoor settings. And epidemiological investigation and traceability efforts should be conducted to prevent and reduce the occurrence of indoor and outdoor clusters. In addition, implementing the joint screening strategy is urgent in the next step of control. Some regions (49) in China have begun to adopt a combined screening strategy of 14-day centralized isolation, 7-day centralized or home isolation and multiple nucleic acid and serum total antibody screening (50, 51) for inbound personnel, which can effectively improve the detection rate and reduce the transmission risk. This study has limitations. First, we only used a limited number of samples. A larger sample might have led to a higher generalization of our results. Second, our research focused on natural history of disease, and it might be important to include other factors as well. Studies (38, (52) (53) (54) have suggested factors that would have an impact on the epidemic situation, such as climate, use of mask, inflow and outflow of population, clinical classification of confirmed cases, reinfection, virus mutation, presymptomatic infection, undiagnosed infections, hospitalization or home isolation measures for confirmed cases. In subsequent studies, consideration of the above factors can be added to make the dynamic model more consistent with the actual situation. In addition, a household-community-factorybased COVID-19 transmission scenario could be constructed based on our calculated transmissibility of the three scenarios. In Chaoyang District, the epidemiological features of SARS-CoV-2 showed multi-stage pattern. Many clusters were reported to occur indoors, mostly from households and factories, and few from the community. The risk of transmission varies by setting, with indoor settings being more severe than outdoor settings. Reported household clusters were the predominant type, but the population size of the different types of clusters limited transmission. The transmissibility of SARS-CoV-2 was different between a single population and its subgroups, with cluster-level transmissibility higher than population-level transmissibility. The raw data supporting the conclusions of this article will be made available by the authors without undue reservation. SY, SC, JR, QL, TC, SW, ZZ, BD, CL, KL, YW, and ZY had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. SW, TC, and QL were responsible for its conception and design, made critical revision of the manuscript for important intellectual content, and contributed equally to the supervisions of this work. SC and SW collected the data. SY, SC, JR, ZZ, BD, CL, KL, YW, and ZY were responsible for the analysis, or interpretation of data. SY, SC, and JR drafted the manuscript. All authors contributed to the article and approved the submitted version. This study was partly supported by the Bill and Melinda Gates Foundation (Grant Number: INV-005834). The funder had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. Comparing SARS-CoV-2 with SARS-CoV and influenza pandemics How COVID-19 Spreads Interim Clinical Guidance for Management of Patients with Confirmed Coronavirus Disease (COVID-19) The epidemiological characteristics of an outbreak of 2019 novel coronavirus diseases (COVID-19) in China Epidemic models of contact tracing: systematic review of transmission studies of severe acute respiratory syndrome and middle east respiratory syndrome Estimate of the basic reproduction number for COVID-19: a systematic review and meta-analysis China coronavirus: what do we know so far? The reproductive number of the Delta variant of SARS-CoV-2 is far higher compared to the ancestral SARS-CoV-2 virus Investigation of a cluster epidemic of COVID-19 in Ningbo Investigation on a cluster epidemic of COVID-19 in a supermarket in Liaocheng, Shandong province Investigation of transmission chain of a cluster COVID-19 cases What settings have been linked to SARS-CoV-2 transmission clusters? Study on transmission dynamic of 15 clusters of COVID-2019 cases in Ningbo Contagiousness and secondary attack rate of 2019 novel coronavirus based on cluster epidemics of COVID-19 in Guangzhou Transmission of SARS-CoV-2: a review of viral, host, and environmental factors Johns Hopkins Bloomberg School of Public Health Cluster infections play important roles in the rapid evolution of COVID-19 transmission: a systematic review Estimating the transmissibility of coronavirus disease 2019 and assessing the effectiveness of the countermeasures to control the disease in Xiamen City A five-compartment model of age-specific transmissibility of SARS-CoV-2 Effectiveness of Interventions to Control Transmission of Reemergent Cases of COVID-19 The estimation of the basic reproduction number for infectious diseases Available online at General Office of National Health Commission, Office of National Administration of Traditional Chinese Medicine. Protocol of diagnosis and treatment for COVID-19(trial version 8) A mathematical model for simulating the phase-based transmissibility of a novel coronavirus Transmissibility of acute haemorrhagic conjunctivitis in small-scale outbreaks in Hunan Province Characteristics of COVID-19 epidemics at different stages in Beijing from The incubation period of 2019-nCoV from publicly reported confirmed cases: estimation and application. medRxiv Incubation Period and Other Epidemiological Characteristics of Novel Coronavirus Infections with Right Truncation: A Statistical Analysis of Publicly Available Case Data Incubation period of 2019 novel coronavirus (2019-nCoV) infections among travellers from Wuhan, China Epidemiological characteristics of 1212 COVID-19 patients in Henan, China. medRxiv A familial cluster of pneumonia associated with the 2019 novel coronavirus indicating personto-person transmission: a study of a family cluster Analysis on cluster cases of COVID-19 in Tianjin Coronavirus disease outbreak in call center COVID-19 Among Workers in Meat and Poultry Processing Facilities-19 Centers for Disease Control and Prevention, Occupational Safety and Health Administration. Meat and Poultry Processing Workers and Employers Sensitivity and specificity of nucleic acid testing in close contacts of COVID-19 cases in Guangzhou Reconstruction of the full transmission dynamics of COVID-19 in Wuhan Evaluation of the secondary transmission pattern and epidemic prediction of COVID-19 in the four metropolitan areas of China First-wave COVID-19 transmissibility and severity in China outside Hubei after control measures, and second-wave scenario planning: a modelling impact assessment Modelling strategies for controlling SARS outbreaks Estimation of MERScoronavirus reproductive number and case fatality rate for the spring 2014 Saudi Arabia outbreak: insights from publicly available data The transmissibility estimation of influenza with early stage data of small-scale outbreaks in Changsha, China Update: COVID-19 among workers in meat and poultry processing facilities-United States Indoor time-microenvironment-activity patterns in seven regions of Europe Transmission of 2019-nCoV infection from an asymptomatic contact in Germany Modeling the transmission dynamics of COVID-19 epidemic: a systematic review Basic reproduction number Effectiveness of the "14 plus 7 day quarantine" and "nucleic acid plus total antibody testing" strategy for screening imported patients with COVID-19 in Xiamen Serology characteristics of SARS-CoV-2 infection after exposure and post-symptom onset Antibody tests for identification of current and past infection with SARS-CoV-2 Fitting and forecasting the trend of COVID-19 by SEIR+CAQdynamic model The impact of meteorological and environmental conditions on the spread of COVID-19 COVID-19 in children in the United States: intensive care admissions, estimated total infected, and projected numbers of severe pediatric cases in 2020 Conflict of Interest: The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest The authors thank the staff of Chaoyang District Center for Disease Control and Prevention for their help in collecting the data.