Supplement for Cloud Computing: Opioid Policy Models

项目总结/摘要 在我们的研究中，北卡罗来纳州的阿片类药物政策建模（5 R 01 DA 04799），我们开发了一个基于代理的 模拟模型（ABM），以告知北卡罗来纳州的政策制定者和卫生专业人员， 全州范围内的干预措施，以减少阿片类药物过量和相关死亡。在国家信息通报中确定了干预措施 阿片类药物行动计划涵盖三大支柱：预防，与护理的联系和减少伤害。 我们的ABM代表一个社区（例如，县、镇）的网络个人（例如，病人，医生， 经销商），并模拟拟议的干预措施如何影响阿片类药物滥用和其他 结果（即，OD死亡）。该模型依赖于一个合成的人口代表每一个人在 美国，它允许多个数据集的信息在一个数据集中概率连接 模型合成个体被定位在地理空间背景中，包括医院的位置， 其他治疗设施和服务提供商。 云基础设施将使我们能够加快现实的模拟，并考虑更多的场景选项， 在当前的“内部”计算基础设施上进行考虑。它将使我们能够开发一个Meta- 模型总结了多种情况。云计算的好处将影响地理空间校准， 验证和情景模拟。

PROJECT SUMMARY/ABSTRACT In our study, Opioid Policy Modeling in North Carolina (5R01DA04799), we developed an agent-based simulation model (ABM) to inform policy makers and health professionals in North Carolina about local and state-wide interventions to reduce opioid overdoses and related deaths. Interventions are identified in the NC Opioid Action Plan and cover the Three Pillars: prevention, connection to care, and harm reduction. Our ABM represents a community (e.g., a county, town) of networked individuals (e.g., patients, physicians, dealers), and simulates how proposed interventions affect individual pathways to opioid misuse and other outcomes (i.e., OD death). The model relies on a synthetic population representing every individual in the United States, which allows information from multiple datasets to be probabilistically connected in one model. Synthetic individuals are positioned in geospatial context that includes the locations of hospitals, other treatment facilities, and service providers. Cloud infrastructure will allow us to speed up realistic simulations and consider more scenario options than are considered on a current “in-house” computational infrastructure. It will enable us to develop a meta- model summarizing a multitude of scenarios. Benefits of cloud computing will affect geospatial calibration, validation, and scenario simulations.