Research On Multi-scale Modeling And Data Analysis Of Individual Behavior Changes And Communication Dynamics

In recent years,emerging infectious diseases coming one after another,such as SARS in 2003,A/H1N1 in 2009,H7N9 in 2013,dengue fever and Ebola in 2014,and annual different pandemic influenza,have brought serious threats to human lives and properties.Therefore,the prevention and control of emerging infectious diseases have been highly concerned by disease prevention and control departments and public health departments for a long time.However,for emerging infectious diseases,due to the sudden outbreak,short transmission period,great risk,and lack of understanding of its transmission mechanism and pathogenesis for humans,it is almost impossible to develop effective drugs and vaccines within a short period.Thus,non-phaxmaceutical interventions for the prevention and control of emerging infectious diseases are particularly important.The effectiveness of such measures depends not only on whether the individual's behavioral habits have changed,but also on air quality,meteorological variables and other factors.Therefore,how to evaluate and assess the effects of individual behavior changes,air quality and mete-orological factors on the prevention and control of emerging infectious diseases have important research value and application prospects.The classical compartment models for infectious diseases could reflect the health states related to transmission(such as in an SIR model:susceptible,infectious and recovered),age or other related individuals' characteristics,but those modelling approach ignored the difference among individuals in the same compartment in terms of infectivity,susceptivity,the level of immunization,and behaviors.Therefore,in order to determine the rate of individual behavior change and then address the effects of individual behavior change rate on the prevention and control of emerging infectious diseases,we develop novel mathematical models,statistical techniques and computational methods by realizing the coupling with population level and individual level based on social networks,and embedding the psychological model of individual decision-making into logistic model(LHBM).In addition,complex multi-scale models coupled with the dynamic development of respiratory infection cases,air quality index(AQI),meteorological variables and individual behavior change are established to reveal the complex relationship between these factors and diseases.Firstly,based on A/H1N1 and Ebola reported cases,combined with media re-lated news reports,multiple individual behavior change models including health beliefs are developed.To obtain the optimal models based on A/H1N1 data,and thus to evaluate the effectivity of individual behavior changes on the whole trend of outbreaks of emerging infectious diseases,we employ the approximate Bayesian computation based on sequential Monte Carlo(ABC SMC)method for model s-election.Besides,we verify the validity of the selected model and evaluate the effects of individual behavior changes on Ebola outbreak.The main results indi-cate that the classical compartment SIR model and the model with average rate of behaviour change depicted by an exponential function could fit the observed data best.However,sensitivity analyses indicate that the accumulated number of hospi-tal notifications and deaths could be largely reduced as the rate of behaviour change increases.Therefore,in terms of mitigating emerging infectious diseases,both media publicity focused on how to guide people's behaviour change and positive responses of individuals are critical for infectious disease control.We develop a multi-scale model by coupling a psychological model of individual decision-making based on social networks with infectious disease model at individ-ual level in the second part.Besides,we formalize individual-based models(IBMs)to incorporate the rate of change of individuals' behavior and investigate how indi-viduals' behaviour changes affect the dynamical evolution of A/H1N1.In addition,ABC SMC is employed to estimate unknown parameters.Furthermore,to obtain the probability distribution of each state during the whole transmission process of infectious diseases,Kolmogorov forward equations were established based on a mod-ified individual-based SIR model.The main results indicate that the novel model developed here can depict the real transmission process and reveal the effects of individual behavior changes on A/H1N1 outbreak more accurately.Moreover,it reveals that it is beneficial to disease prevention and control for media publicity on how to guide individuals' behaviour changes.We conduct an integrated data analysis to quantify the association among AQI,meteorological variables and influenza-like illness(ILI)cases in Shaanxi province in the third part.To realize parameter estimation and variable selection correspond-ing to high-dimensional data of air pollution monitoring indicators,meteorological variables and ILI cases,generalized additive model is established and relative risk analysis is conducted.Thus the core factors affecting ILI cases are precisely de-termined among them.Our analysis illustrates a statistically significantly positive correlation between the number of ILI cases and AQI,and the respiratory infection risk has increased progressively with increased AQI with a time lag of 0-3 days.We also develop mathematical models for the AQI trend and respiratory infection dynamics,incorporating AQI-dependent incidence and AQI-based behaviour change interventions.Our combined data and modelling analysis allow us to estimate the basic reproduction number for the respiratory infection during the studying period which is given as 2.4076,higher than the basic reproduction number of the 2009 A/H1N1 influenza in Shaanxi Provice,which indicates the significant effect of smog on respiratory diseases.Sensitivity analyses reveal the impact of air pollution on ILI cases and could contribute to evaluate the effectiveness of control strategies for air pollution,so as to provide qualitative and quantitative decision-making basis for air pollution and respiratory disease control.The main conclusions indicate that,in terms of respiratory infection risk reduction,the persistent control of emission in the China's blue-sky programme is much more effective than substantial social-economic interventions implemented only during the smog days.