| Acute respiratory infectious diseases such as COVID-19 and influenza are one of the greatest threats to human survival and have affected the development process of human civilization several times.Building a mathematical model to simulate the spread of respiratory infectious diseases is an effective way to understand the key factors that influence the spread of infectious diseases,predict infectious disease trends,and scientifically formulate various prevention and control measures.It has been widely used in epidemic risk assessment and scientific decision-making in the formulation of intervention plans.For megacities with dense populations,developed economies,intensive employment,and active leisure and entertainment activities,the spread of respiratory infectious diseases has been promoted and expanded,and the pressure for prevention and control is enormous.Therefore,it is particularly important to construct a spatiotemporal diffusion model of acute respiratory infectious disease epidemics（"epidemics" for short）within cities to assist in the scientific formulation of urban refined prevention and control strategies.The temporal and spatial evolution of urban population activities and the associated fluctuations in population density are the key factors in the spread of epidemics in cities.The heterogeneity and dynamics of interactions between urban individuals are particularly important for the spread and control of the epidemic.Individual demographic attributes and interactions among individuals are heterogeneous in the agent-based epidemic diffusion model compared with the meta-population epidemic diffusion model,which can help simulate the randomness of the nonlinear system of infectious disease transmission at different spatial scales.However,building an agent-based epidemic diffusion model requires a large amount of real individual data.Traditional agent-based epidemic diffusion models could not accurately simulate individual locations and contacts between individuals due to a lack of large-scale individual trajectory data.The explosive growth of individual trajectory data provides a new opportunity to reconstruct accurate individual travel locations.Although some agent-based epidemic diffusion models integrate individual trajectory data to simulate urban population mobility,the existing modeling methods cannot effectively integrate population attributes and multi-source trajectory movement characteristics for individual modeling and lack effective methods to construct the complexity of urban individual space-time contact relationships.In addition,the current agent-based epidemic diffusion model lacks a refined modeling method and application of intervention measures at the individual and building levels.At the same time,the method for recommending intervention measures based on the main principle of infection size also lacks a control concept that takes into account the risk of epidemic outbreaks.To overcome the above problems and face the refined scale of the city,this study integrates large-scale trajectory data to construct an agent-based epidemic spatio-temporal transmission model and applies it to the spread simulation,transmission characteristics analysis,and intervention recommendations of COVID-19 and influenza in Shenzhen.Specifically,this research first integrates multi-source trajectory data and various types of urban spatio-temporal big data to construct a dynamic contact network of urban whole-population agents.Then,taking the two typical acute respiratory infectious diseases,COVID-19 and influenza,as examples,a refined simulation model of epidemic spatio-temporal spread and epidemic intervention measures at the level of "individual-building-activity type" was constructed.With the support of the super computing environment,the epidemic development in different scenarios was deduced,the accuracy of the model was verified based on the epidemic curve,age distribution of infected people,and other dimensions at multiple spatial scales within the city,and some spatio-temporal transmission characteristics of related epidemics in the city were explored.Finally,on the one hand,in view of the transmission characteristics of the original strain of COVID-19,in order to suppress the second outbreak of the COVID-19 epidemic in megacities,a combination of non-pharmaceutical interventions that take into account the probability of outbreaks is proposed based on the simulation results of the COVID-19 model.On the other hand,based on some effective prevention and control measures developed in the COVID-19 epidemic,combined with the key transmission characteristics discovered by the influenza model,a combination of influenza epidemic interventions for the post-COVID-19 era was systematically evaluated and recommended.The contributions and innovations of this research include:First,large-scale mobile phone location data and travel survey data are integrated to construct a dynamic contact network of urban whole-population agents at the level of "individual-building-activity type" to reduce the spatial uncertainty of the simulation of the epidemic spread process.To address the issue that the previous agent-based epidemic spread model could not accurately simulate individual location and contact,synthetic individuals equal to the total population of the study area are synthesized by integrating large-scale trajectory data and multi-source urban big data,and population attributes and individual activity chains are assigned to each individual.The locations in each activity chain are represented by buildings,and there are four activity types: home,school,work,and leisure.According to their activity chain,24spatio-temporal co-occurrence networks are created for each individual per day,and then the individual’s regular contact and casual contact are simulated.Finally,this study generates a dynamic contact network for 12 million individuals.Second,in order to suppress the secondary outbreak of the COVID-19 pandemic in megacities during the first wave of the COVID-19 epidemic,a combination of non-pharmaceutical interventions considering the outbreak probability was proposed based on an agent-based epidemic spatiotemporal diffusion model supported by trajectory data.In terms of the transmission characteristics of the original strain of COVID-19,and in the absence of COVID-19 vaccines and specific drugs,this study evaluated the probability of COVID-19 resurgence if sporadic cases occurred under different combinations of implementation intensities of three measures: contact tracing,mask wearing,and prompt testing.In order to reduce the probability of a COVID-19 sporadic epidemic outbreak to under 5% within four weeks,the strategy of "tracking all close contacts + 80% compliance with mask use + 40% prompt testing" is recommended as the lowest prevention and control level in Shenzhen and other megacities.At the same time,the corresponding combination of intervention measures for cities with different anti-epidemic abilities is systematically recommended.Third,Based on some effective prevention and control measures developed during the prevention and control of the COVID-19 epidemic,a combination of influenza epidemic intervention measures suitable for the post-COVID-19 epidemic era was systematically evaluated and recommended.Since the outbreak of the COVID-19 epidemic,public awareness of vaccinations,mask wearing,and home quarantine after onset has generally increased.In order to control the influenza epidemic and reduce the proportion of infected people with a high mortality rate,this study simulated the relationship between the infection size of an influenza epidemic and different combinations of vaccination,mask-wearing,and home-quarantine rate without considering strict intervention measures.Taking Shenzhen as an example,the simulation results showed that adults were a high-risk population with a low reporting rate,and children formed the lowest infected proportion and had the highest reporting rate;the workplace is the second contributor to influenza transmission other than family.Based on the above influenza transmission characteristics,the model found that 45% of adults continue to be vaccinated after influenza vaccination for the elderly and children,while ensuring that at least 60% of infected adults wear masks and at least 20% of infected adults home quarantine after the onset of the disease can significantly reduce the infection scale of influenza epidemics in cities and the infection rate of children and the elderly with high mortality in influenza epidemics.By integrating mobile phone location data,this study more accurately simulates individual mobility and contact networks,develops a new modeling and simulation method of respiratory infectious disease diffusion on the urban scale,and recommends a combination of refined interventions for the actual prevention and control of the COVID-19 and influenza epidemics,which provides corresponding method support and application exploration for in-depth revealing the spatiotemporal transmission characteristics of the epidemic and scientifically recommends accurate intervention measures. |
PDF Full Text Request