Collaborative Research: RAPID: Behavioral Epidemic Modeling For COVID-19 Containment

在流行病期间，人们往往会改变和调整自己的行为。从历史上看，行为适应在流行病的发展中发挥了核心作用。在某些情况下，行为的改变可能会抑制流行病(例如，通过恐惧驱使的逃亡或自我隔离)，而在其他情况下，行为的改变可能会加剧疾病的传播(例如，通过拒绝接种疫苗或过早停止疏远)。在当前的新冠肺炎疫情中，人们未来的行为将决定我们如何应对两个核心威胁。首先是由于过早取消社会疏远指导方针以及很大比例的人口放弃社会疏远而造成的接二连三的疫情浪潮的直接威胁。第二个威胁是，即使在疫苗可用后，疾病仍将反弹。这种情况以前曾发生过(例如麻疹)，如果足够一部分人口因恐惧而拒绝接种疫苗，新冠肺炎也可能发生这种情况。这项研究将开发一种新的模型，使用公开的社交媒体数据来预测人类的行为，以解决大多数现有模型中的一个已知弱点。新的模型、源代码、数据、参数、假设和方法都将完全开放和公开，确保所有结果的可复制性。该项目还将提供该模型的互动版本，供政策制定者、政府机构和教育工作者使用。其他更广泛的影响是研究生的培训机会。目前用于预测流行病进程的大多数模型可能提供不完整的结果，因为它们没有考虑到人们行为的变化(人类行为适应)。在这项拟议的研究中，人类行为将被包括在一个新的模型中，该模型使用从Twitter和Facebook等社交媒体平台获取的数据。这些数据将允许对该模型进行校准，并复制迄今为止整个纽约州的疫情。机器学习方法将被应用于在广泛的情景和控制策略中确定最佳信息和干预措施。这一发展需要一个独特的跨学科团队，涵盖社会科学、传染病建模、生物统计学、社交媒体数据挖掘和机器学习。这个快速奖项由环境生物学部门的传染病生态学和进化计划颁发，使用冠状病毒援助、救济和经济安全(CARE)法案的资金。该奖项反映了NSF的法定使命，并通过使用基金会的智力优势和更广泛的影响审查标准进行评估，被认为值得支持。

People tend to change and adapt their behavior during epidemics. Historically, behavioral adaptation has played a central role in the progression of epidemics. In some cases, a change in behavior may suppress an epidemic (e.g., through fear-driven flight or self-isolation) while in other cases, a change may intensify the spread of a disease (e.g., through vaccine refusal or premature cessation of distancing). People’s future behavior during the current COVID-19 epidemic will determine how well we cope with two central threats. First is the immediate threat of successive epidemic waves due to the premature lifting of social distancing guidelines, and the abandonment of social distancing by a large percentage of the population. The second threat is that the disease will rebound even after a vaccine is available. This has occurred before (e.g., measles) and could occur with COVID-19 if a sufficient fraction of the population refuses vaccine out of fear. This research will develop a new model to predict human behavior using publicly available social media data to address a known weakness in most current models. The new model, source code, data, parameters, assumptions, and methods will all be completely open and publicly available, ensuring the replicability of all results. This project will also deliver an interactive version of the model for use by policymakers, government agencies, and educators. Other broader impacts are training opportunities for a graduate student. Most current models that are used to forecast the course of epidemics may provide incomplete results because they do not take into account changes in peoples’ behavior (human behavioral adaptation). In the proposed research, human behavior will be included within a new model using data taken from social media platforms such as Twitter and Facebook. The data will allow calibration of the model and replicate the entire New York State epidemic to date. Machine Learning approaches will be applied to determine optimal messages and interventions over a wide range of scenarios and control strategies. This development requires a unique interdisciplinary team spanning social science, infectious disease modeling, biostatistics, social media data mining, and Machine Learning.This RAPID award is made by the Ecology and Evolution of Infectious Diseases Program in the Division of Environmental Biology, using funds from the Coronavirus Aid, Relief, and Economic Security (CARES) Act.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.