Behavioral Epidemiology: Applications of Agent-Based Modeling to Infectious Disea

摘要： 模型在包括大流行流感在内的重要公共卫生领域发挥着核心作用， 慢性病和备灾。然而，行为因素是 在当前的建模中几乎被忽视了，这是一个关键的缺陷。行为可以 从根本上影响传染病(如流感)的传播。人民 可能藐视或不知道政府的遏制指令。不信任 社区可能会拒绝接种疫苗。而不是自我孤立、暴露的个人 可能会在恐惧中逃离，加速疾病的空间传播。流行病建模以 Date几乎忽略了这些行为适应及其后果。 行为也会影响慢性病的结果--但没有任何行为机制是这样的 目前被用来解释，例如，引人注目的历史动态 肥胖。理论和实证工作可以确定这种机制， 包括同龄人效应。在灾难中，行为也会影响健康结果。 在城市有毒羽流的释放中，自然的逃离冲动可能会放大 拥堵，破坏疏散，增加暴露。有没有简单的 去中心化的规则或定制的消息，相反，可以产生高效的 疏散？压力下的行为可能基于糟糕的信息， 毫无根据的恐惧，或根深蒂固的社会规范。这些可能性应该是 条件风险沟通与遏制/预防的设计 各种公共卫生威胁的战略。目前，他们并没有这样做。 因此，这项研究的一个中心目标是建立有界理性的内生模型 时空疾病的行为适应及其反馈效应 动力学。该项目涵盖了行为的流行病学(如恐慌、 不遵守、不信任)和行为因素在进展中的作用 传染病和慢性病。 核心分析技术将是基于代理的建模，通过多个 学科研究团队。PI是这一创新领域公认的先驱 曾指导过成功的多学科代理建模项目 天花、考古学、经济学和民间暴力，正如他在书中所述， 生成社会科学：基于主体的计算建模研究 (普林斯顿，2006)。

Abstract: Models are central in important areas of public health, including pandemic flu, chronic disease, and disaster preparedness. However, behavioral factors are virtually ignored in current modeling, a crucial defect. Behavior can fundamentally shape the spread of infectious diseases (such as influenza). People may flout, or not know, government containment directives. Distrustful communities may refuse vaccine. Rather than self-isolate, exposed individuals may flee in fear, accelerating the spatial spread of disease. Epidemic modeling to date has virtually ignored these behavioral adaptations and their consequences. Behavior also shapes chronic disease outcomes-yet no behavioral mechanism is currently offered to explain, for example, the striking historical dynamics of obesity. Theoretical and empirical work could identify such mechanisms, including peer effects. Behavior would also shape health outcomes in a disaster. In an urban toxic plume release, the natural impulse to flee could amplify congestion, undermining evacuation, and increasing exposure. Are there simple decentralized rules or tailored messages that could, instead, generate efficient evacuation? Behavior under stress may be based on poor information, unwarranted fear, or entrenched social norms. These possibilities should condition risk communication and the design of containment/preventive strategies across the spectrum of public health threats. At present, they do not. A central goal of this research, then, is to model boundedly rational endogenous behavioral adaptations and their feedback effects on spatio-temporal disease dynamics. The Project covers both the epidemiology of behaviors (such as panic, non-compliance, distrust) and the role of behavioral factors in the progress of infectious and chronic diseases. The core analytical technique will be agent-based modeling by multi- disciplinary research teams. The PI is a recognized pioneer in this innovative field, and has directed successful multi-disciplinary agent modeling projects on Smallpox, Archaeology, Economics, and Civil Violence, as recounted in his book, Generative Social Science: Studies in Agent-Based Computational Modeling (Princeton, 2006).