SCH: INT: AURA - Connecting Audio and Radio Sensing Systems to lmprove Care at Home

从医院到家庭的有效护理过渡对于确保高危患者获得最佳护理至关重要 并降低可避免的再入院的成本，从而减轻医疗保健系统的负担。常见做法 技术辅助护理转型的目的是通过网络培训和指导患者和家庭护理人员 以及基于移动电话的系统，用于症状报告、生命体征监测和提供反馈。 虽然有效，但这些系统具有三个主要局限性：(a) 患者或护理人员必须主动 测量数据并将其输入系统——这容易出错、主观且经常被遗忘； (b) 网络或 基于移动设备的交互可能是繁琐且要求较高的任务，例如输入数据、了解健康状况 状态，或获取和响应警报需要输入并单击一系列表单/网络 页面； (c) 系统通过消息和通知提供的反馈往往无效且 未被注意到。为了克服这些限制，该项目采用数据驱动的方法来开发 家庭环境中支持语音、情境感知的治疗后自我护理系统。拟议的系统 (a) 使用先进的基于 WiFi 无线电信号的人体监测患者在家中的具体活动 活动识别算法； (b) 定制亚马逊等语音助手的自然语言响应 Echo/Alexa 基于患者的位置、活动和健康史； (c) 自动化数据输入和 报告以减轻患者或护理人员的负担。该系统将部署在 40 个结直肠和膀胱 癌症患者之家评估可用性、可行性和初步效益程度。 该项目的总体目标与 NLM 2017-2027 年的 10 年战略计划一致。 通过语音和 WiFi 无线电传感器数据的自动收集、链接、管理和建模，这 项目将创建一个相互关联的数字生物标记生态系统，可以解释、影响和预测 护理过渡期间的健康结果。通过使用低成本的语音助手（低于 50 美元）和无处不在的 家庭 WiFi，它最大限度地向大众传播和参与数据驱动的健康。 为了培养未来数据就绪的劳动力队伍，三名博士。学生将从事此工作 多学科项目，将开发一门新的移动医疗课程，并将开放科学实践 在项目的开发、部署和传播阶段应用。

An effective care transition from hospital to home is crucial to ensure optimal care for high-risk patients and to reduce the cost of avoidable readmissions that burdens the healthcare system. A common practice in technology-assisted care transition is to train and guide patients and family caregivers through the web and mobile phone-based systems for symptom reporting, vital signs monitoring, and providing feedback. Although effective, these systems have three major limitations: (a) the patient or caregiver has to actively measure and enter data into the system-which is error-prone, subjective, and often forgotten; (b) web or mobile-based interactions can be cumbersome and demanding-tasks like entering data, knowing health status, or getting and responding to an alert require typing and clicking through a series of forms/web pages; and (c) feedback from the system through messages and notifications is often ineffective and unnoticed. To overcome these limitations, this project employs a data-driven approach to develop a voice-enabled, context-aware, post-treatment self-care system in home settings. The proposed system will (a) monitor specific activities of a patient at home using advanced WiFi radio signal-based human activity recognition algorithms; (b) tailor natural language responses of voice assistants like Amazon Echo/Alexa based on the patient's location, activity, and health history; and (c) automate data-entry and reporting to reduce patient or caregiver burden. The system will be deployed in 40 colorectal and bladder cancer patients' homes for assessing usability, feasibility, and preliminary magnitude of benefits. The overarching goals of this project are aligned with the NLM's 10-year strategic plan for 2017-2027. Through automated collection, linking, curation, and modeling of voice and WiFi radio sensor data, this project will create an interconnected ecosystem of digital biomarkers that explain, influence, and predict health outcomes during care transition. By using low-cost voice assistants (less than $50) and ubiquitous home WiFi, it maximizes the dissemination and engagement of data-powered health to mass population. To foster the development of a data-ready workforce for the future, three Ph.D. students will work in this multidisciplinary project, a new mHealth course will be developed, and open science practices will be applied during the development, deployment, and dissemination phases of the project.