SCC-PG: Using Innovations in Sensing, Data Analytics, and Community Engagement to Address Opioid Overdose Crisis

阿片类药物过量现在是美国50岁以下人群死亡的主要原因。各城市采取了不同的战略，通过各种教育/培训方案解决这一问题。然而，美国阿片类药物过量危机的规模越来越大，这表明需要更有效的数据驱动方法。阿片类药物滥用和过量已被确定为弗吉尼亚州里士满地区过早死亡的主要方法。数据不足是市政府官员面临的一个主要问题，这使他们无法调查阿片类药物流行的规模。由于地方的经济竞争力及其招募企业和劳动力的能力取决于它们所描绘的形象，因此阿片类药物过量的数据共享和进一步分析的改善有限。为此，该规划项目建议建立一个由研究人员和当地利益攸关方组成的团队，包括里士满地区所有城市和县的领导人员。该团队将努力以数据为导向理解问题，并提出社区参与的解决方案来解决问题。由于对阿片类药物问题的反应在美国其他大都市地区很常见，我们希望在其他地区推广这一模式。该项目旨在利用数据分析和智能技术的力量，为有效的决策和规划开发创造性的解决方案，以改善公共卫生和生活水平。更广泛的直接影响包括增强和削弱社区对阿片类药物流行病的反应的循证因素，这些因素可以与社区成员进行讨论和改进，以推动变革，从而减少整个地区的阿片类药物过量和健康不公平现象。从技术角度来看，该项目将研究新的数据驱动的治疗政策方法，这些方法可以得到社区的支持。这项工作的智力价值包括：（1）从根本上了解社区因阿片类药物流行病而面临的挑战，（2）更好地了解治理，智慧城市和社会创新之间的关系，特别是解决阿片类药物问题，（3）收集社区中相关类型的药物使用数据，（4）根据来自不同来源的可用数据导出预测模型，以及（5）开发智能传感解决方案，以准确监测和评估药物滥用状况。因此，该项目旨在开展跨学科努力，制定数据驱动的干预措施，与社区代表协调，确定和评估适用的方法，以解决阿片类药物过量危机。该小组将通过考虑与毒品有关的事件的各种数据，调查用于预测毒品使用/过量的几个预测模型。该项目还将研究可解释的机器学习技术，这些技术将与开发的数据驱动模型相结合，以提供对药物使用的估计，并解释证明模型预测合理的重要因素。该奖项反映了NSF的法定使命，并通过使用基金会的知识价值和更广泛的影响审查标准进行评估，被认为值得支持。

Opioid overdose is now the leading cause of death for those under 50 in the USA. Cities have followed different strategies to address this problem through various education/training programs. However, the growing scale of the opioid overdose crisis in the USA indicates that more effective data-driven approaches are needed. Opioid abuse and overdose have been identified as a leading method of premature death in the Richmond Region, Virginia. Inadequate data is a major issue for city officials, which prevents them from investigating the scale of the opioid epidemic. Since locality’s economic competitiveness and their ability to recruit businesses and labor are dependent on the image they portray, the data sharing and further analytics in opioid overdoses have seen limited improvement. To this end, this planning project proposes to build a team of researchers and local stakeholders - including those in the leadership roles in all cities and counties in the Richmond Region. The team will work towards a data-driven understanding of the problem and community-involved solutions to address the issue. As responses to the opioid problem are common to other metro regions in the U.S., we hope to scale the model to be exercised in other regions. This project aims to harness the power of data analytics and smart technologies to develop creative solutions for efficient decision-making and planning to improve public health and living standards. Direct broader impacts include evidence-based factors that enhance and impair community responses to the opioid epidemic that can be discussed and refined with community members to drive change, and accordingly, reduce opioid overdose and health inequities across the region.From the technical perspectives, this project will investigate novel data-driven approaches to treatment policies that can be supported by the community. The intellectual merit of this work includes: (1) developing a fundamental understanding of challenges facing communities due to opioid epidemics, (2) developing a better understanding of the relationship between governance, smart cities, and social innovation, particularly for addressing the opioid problem, (3) collecting relevant types of drug use data in the community, (4) deriving prediction models based on the available data from different sources, and (5) developing smart sensing solutions to accurately monitor and assess the state of drug abuse. Accordingly, this project aims to establish interdisciplinary efforts to develop a data-driven intervention in addressing the opioid overdose crisis, in coordination with community representatives to identify and assess applicable approaches. The team will investigate several predictive models for forecasting drug use/overdoses by considering diverse data on drug-related incidents. The project will also investigate Explainable Machine Learning techniques that will be coupled with developed data-driven models in order to provide estimations of drug use with an explanation of the important factors that justify the predictions made by the model. This will help identify the root causes and the extent of their impact.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.