The diffusion of influenza disease and that of individual preventive behavior are intrinsically interrelated. This dissertation presents an original dual-diffusion model to couple these two diffusion processes together. At the conceptual level, the model is composed of four components, including the contact network, diffusion of influenza, diffusion of preventive behavior, and the interaction between the two diffusion processes. The individual-based approach, network structure, disease model and threshold adoption model have been integrated under this framework to formulate each component. The model implementation is carried out in the urbanized area of Buffalo, New York, based on census data, land use data, travel statistics, and health behavior survey. The dual-diffusion model successfully replicates the observed trends of influenza infection and antiviral-drug use, and offers a close representation of the lab-confirmed data as well.
The presented model is further used to investigate spatial-temporal dynamics of the dual-diffusion processes. The simulation results identify five subsequent stages during the course of dual-diffusion, including: the local growth, expansion, fast city-wide growth, slow city-wide growth, and fade-out. Two major factors are found to contribute to the dynamics. One is the spatial heterogeneity in the city, in terms of the population distribution and land use patterns. This factor directly influences the spatial layout of the dual-diffusion. The other factor is the travel of individuals in the city, which has profound effects on the temporal sequence of the dual-diffusion.
One purpose of developing the dual-diffusion model is to explore effective strategies for influenza control and prevention. This research evaluates the combined effectiveness of control strategies and individual preventive behavior. Three control strategies and their combination have been simulated, including a targeted antiviral prophylaxis (TAP) strategy, a workplace closure strategy, and a travel restriction strategy. The results suggest that previous studies on control strategies may have under-estimated the resultant effectiveness, because the effects of preventive behavior are overlooked. The addition of preventive behavior may double, or even triple, the control effectiveness, leading to smaller disease attack rates and lower epidemic peaks. The control strategies previously suggested might be resource-intensive, and optimized strategies are recommended.
In addition to the control strategies, two preventive strategies have been simulated and assessed: one strategy offers free antiviral drugs to households (referred to as incentive strategy), and the other establishes role models of adoption at affected workplaces (called role-model strategy). Different from previous studies, the assessment not only considers the effects of these strategies on the number of adoptions, but also their effects on the number of infections. The results show that the incentive strategy can be effective to control influenza, given a moderate compliance of households. The role-model strategy is not recommended for the study area, because half of the population may be reluctant to adopt unless they were ill, even if role models are setting around them.
This dissertation expands current epidemic models into a new field, human preventive behavior against diseases. It argues that the interactions between human and disease are reciprocal, and the failure to consider either side may affect decision making. The results offer in-depth understandings in influenza transmission, and control, and prevention. A number of strategies explored in this research can be valuable for the public to overcome socio-economic challenges posed by future influenza epidemics.