SCH: Improving Patient Health and Equity through the Digital Transformation of Diabetes Care Delivery

这一智能和互联健康（SCH）奖将通过发展利用数字技术提高糖尿病管理质量和降低成本的科学基础，为提高国民健康和福利做出贡献。越来越多地采用远程活动追踪器、连续血糖监测仪、其他类型的传感器和远程医疗（由COVID-19大流行促成），改变了糖尿病护理。典型的糖尿病护理标准是每天进行几次“手指戳”血糖测量，每年与护理团队进行几次门诊检查，从而错过了许多检测和纠正血糖管理恶化的机会。糖尿病护理的数字化具有按需测量、监测和个性化患者护理的潜力。在最好的情况下，这些技术进步有望通过向最需要的人提供有限的保健资源，缩小获得保健服务方面的差距。实现这一结果需要新的科学进展来利用传感器数据更好地了解患者；开发算法技术，有效地为患者分配护理资源；并帮助提供者和患者共同制定和实施护理决策。该奖项专门开发算法、平台和决策支持工具，通过跨学科的努力将工程师、计算机科学家和统计学家与临床医生联系起来，解决儿科人群1型糖尿病护理背景下的这些挑战。该项目的一个关键组成部分包括在两个地点（斯坦福大学的露西尔·帕卡德儿童医院和堪萨斯城的儿童慈善医院）进行部署，以捕捉理论必须解决的现实世界的限制，并验证科学创新开发的现实世界的有效性。研究目标是：(1)开发从传感器数据中识别患者类型的方法，为医疗团队了解患者群体的差异化需求和行为提供信息；(2)制定方法，将稀缺的提供者资源分配给最需要护理的患者；(3)设计可解释的治疗计划建议，以促进提供者和患者之间的互动，包括面向提供者和面向患者的“仪表板”，使进展和疾病管理可视化，并指导干预措施的讨论。这些重点既利用并推进了时间序列建模和机器学习的最新技术；行为心理学;以及数据驱动的运筹学。来自临床环境的数据和现实世界的理解有助于定义理论必须解决的问题，并作为验证所开发的模型和方法的基准。开发的模型将在治疗1型糖尿病儿科患者的诊所中进行部署和评估。该奖项反映了美国国家科学基金会的法定使命，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。

This Smart and Connected Health (SCH) award will contribute to the advancement of the national health and welfare by developing the scientific foundation of using digital technology to improve the quality and reduce the cost of diabetes management. Increased adoption of remote activity trackers, continuous glucose monitors, other types of sensors, and telehealth (precipitated by the COVID-19 pandemic) have transformed diabetes care. The typical standard of care for diabetes revolves around a few “finger poke” glucose measurements per day and a few in-clinic visits with the care team per year, with many missed opportunities to detect and remedy deteriorating glucose management. The digitization of diabetes care holds the potential for as-needed measurement, monitoring, and personalized patient care. In a best case scenario, these technological advances promise to narrow gaps in health care access, by providing limited care resources to those most in need. Achieving this outcome requires novel scientific progress to leverage sensor data to better understand patients; to develop algorithmic techniques to allocate care resources efficiently to patients; and to help both providers and patients develop and implement care decisions together. This award specifically develops algorithms, platforms, and decision support tools to address these challenges in the context of care of type 1 diabetes in pediatric populations, by connecting engineers, computer scientists, and statisticians with clinicians in an interdisciplinary effort. A key component of the project involves deployment at two sites (Lucile Packard Children’s Hospital at Stanford, and Children’s Mercy Hospital at Kansas City) to capture the real-world constraints the theory must tackle and to validate the real-world efficacy of the scientific innovations developed.The research objectives are to: (1) develop methods to identify patient types from sensor data, informing the health care team's understanding of differentiated needs and behaviors across the patient population; (2) develop methods to allocate scarce provider resources to those patients most in need of care; and (3) design interpretable treatment planning recommendations to facilitate interactions between providers and patients, including both provider-facing and patient-facing “dashboards” that visualize progress and disease management, and guide discussion of interventions. These thrusts both leverage and advance the state of the art in time-series modeling and machine learning; behavioral psychology; and data-driven operations research. Data and real-world understanding from the clinical context serve to define the problems the theory must tackle and as a benchmark to validate the models and methods developed. The models developed will be deployed and evaluated in clinics caring for pediatric patients with type 1 diabetes.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.