We develop an agent-based model of vaccine decisions across a heterogeneous network model with urban and rural regions. In the model, agents make rational decisions to vaccinate or not, based on the relative private costs of vaccinations and infections as well as an estimated probability of infection if not vaccinated. The model is a methodological advance in that it provides an economic rationale for traditional threshold models of vaccine decision-making that are commonly used in agent-based network models of vaccine choice. In the model, more dense urban regions have more connections between agents than less dense rural regions. Higher density leads to higher levels of vaccine usage and lower rates of infection in urban regions within the model. This finding adds to the more commonly discussed socio-economic reasons for higher levels of vaccination usage in urban areas compared to rural areas. In addition to this direct contribution, the paper emphasizes the importance of endogenous decision-making in models of epidemiology. For instance, we find that networks that lead to larger epidemics in exogenous vaccination models lead to smaller epidemics in our model because agents use vaccinations to offset the additional risk introduced by these network structures. Endogenous agent responses to risk need to be incorporated into theoretical and empirical models of economic epidemiology.

我们在具有城市和农村地区的异构网络模型中开发了基于主体的疫苗决策模型。在该模型中，代理人根据疫苗接种和感染的相对私人成本以及如果不接种疫苗估计的感染概率，做出是否接种疫苗的理性决定。该模型是方法论上的进步，因为它为传统的疫苗决策阈值模型提供了经济原理，这些模型通常用于基于代理的疫苗选择网络模型。在该模型中，人口稠密的城市地区比人口密度较低的农村地区在代理人之间的联系更多。在模型中，较高的密度会导致更高水平的疫苗使用和较低的城市地区感染率。与农村地区相比，这一发现增加了城市地区疫苗接种率更高的更普遍讨论的社会经济原因。除了这一直接贡献外，本文还强调了内生决策在流行病学模型中的重要性。例如，我们发现在外源疫苗接种模型中导致更大流行病的网络会导致我们模型中更小的流行病，因为代理人使用疫苗接种来抵消这些网络结构引入的额外风险。内生代理人对风险的反应需要纳入经济流行病学的理论和实证模型。我们发现在外源疫苗接种模型中导致更大流行病的网络会导致我们模型中更小的流行病，因为代理人使用疫苗接种来抵消这些网络结构引入的额外风险。内生代理人对风险的反应需要纳入经济流行病学的理论和实证模型。我们发现在外源疫苗接种模型中导致更大流行病的网络会导致我们模型中更小的流行病，因为代理人使用疫苗接种来抵消这些网络结构引入的额外风险。内生代理人对风险的反应需要纳入经济流行病学的理论和实证模型。