A system to detect fall occurrence and location in hospital settings

项目总结/摘要 跌倒检测、评估和预防对于医疗机构来说是一个具有挑战性的问题。尽管 跌倒相关损伤是最常见的不良事件，被广泛认为是一个严重的问题 在医院住院环境中|高达20%的患者在住院期间至少跌倒一次。据该机构称 根据医疗保健研究和质量报告，30%至51%的福尔斯跌倒导致受伤，而这些受伤中高达44 导致死亡。仅在美国，每年住院福尔斯的人数就超过100万，其中老年人 个人是最脆弱的。作为其重要性的证明，秋季研究被列为首要任务 在医学研究所向国会提交的关于国家优先事项的报告中。 多种因素导致福尔斯：患者因素（例如，认知状态、肌无力、药物）， 环境因素（例如，床/厕所定位，绊倒危险），和制度因素（例如，不良反应- 对床/椅警报的敏感性，跌倒评估和预防措施不足）。虽然医院采用 预防福尔斯的计划，跌倒预防计划的一个重要组成部分是准确确定 跌倒事件，因为许多福尔斯是“看不见的”和未报告的。在其2016年索赔报告中，保险公司CNA 指出，66%的福尔斯跌倒是无人目击的，并指出在超过一半的已结案的法律的索赔中，未能监测福尔斯跌倒。 替代的跌倒检测解决方案尚未在医院环境中采用，这是由于跌倒灵敏度差、高 虚警率高，缺乏必要的配套功能。 解决方案本文提出的身体佩戴的FallCall系统将利用强大的新颖感测能力， 提供鲁棒的跌倒检测并报告跌倒位置，而没有令人烦恼的错误警报。The FallCall紧身衣 仪器将包含封装在一次性低过敏性 贴片贴片可以放置在躯干或上臂的任何地方，以最大限度地减少与 其他医疗设备或受伤部位。该系统将持续监测患者并及时发出警报 如果检测到跌倒，尽量减少患者在地面上的时间，以及受伤评估和治疗时间。 FallCall系统将允许确定患者位置（包括医院的房间和房间） 使用可靠的、健壮的无线网状网络。此外，FallCall系统将能够发出 徘徊和不活动警报。拟议的系统将是独特的，在其能力，客观地衡量反应 跌倒发生时的临床阶段时间，并将成为评估和改善跌倒预防的关键组成部分 程序. 第一阶段的结果第一阶段的任务是在秋季实验室和 真实世界的医院环境中使用的病人。研究结果表明，100%的福尔斯跌倒检测结果为零假警报。 定位也达到了完美的精度。进行了正式的可用性研究， 根据标准数据，排名前5%。

Project Summary/Abstract Problem Fall detection, assessment, and prevention is a challenging issue for healthcare institutions. Despite wide recognition as a considerable problem, fall-related injuries are the most frequently reported adverse events in hospital inpatient settings|up to 20% of patients fall at least once during their stay. According to the Agency for Healthcare Research and Quality reports, 30% to 51% of falls result in injury, and up to 44% of those injuries lead to death. In the United States alone, the number of inpatient falls exceeds 1 million per year, with elderly individuals being the most vulnerable. As a testament to its importance, fall research was listed as a top priority in the Institute of Medicine's report to Congress on national priorities. Diverse factors contribute to falls: patient factors (e.g., cognitive state, muscle weakness, medications), environmental factors (e.g., bed/toilet positioning, trip hazards), and institutional factors (e.g., poor respon- siveness to bed/chair alarms, inadequate fall assessment and prevention measures). Although hospitals employ programs to prevent falls, an essential component of a fall prevention program is accurate determination of fall events, as many falls are \unseen" and unreported. In their 2016 Claims Report, insurance company CNA stated that 66% of falls were unwitnessed and cited failure to monitor for falls in over half of closed legal claims. Alternative fall detection solutions have not been adopted in hospital settings due to poor fall sensitivity, high false-alarm rates, and lack of necessary supporting functions. Solution The body-worn FallCall system, proposed herein, will leverage powerful, novel sensing capabilities to provide robust fall detection and to report fall location without bothersome false alarms. The FallCall body-worn instrument will embody a miniature electronics package that is encapsulated in a disposable, hypoallergenic adhesive patch. The patch can be placed anywhere on the torso or upper arms to minimize interaction with other medical equipment or injury sites. The system will continuously monitor patients and issue a timely alert if a fall is detected, minimizing patient time on the ground, as well as injury assessment and treatment latencies. The FallCall system will allow patient location to be determined (including the room and oor of the hospital) using a reliable, robust wireless mesh network. Additionally, the FallCall system will have the capability to issue wander and inactivity alerts. The proposed system will be unique in its ability to objectively measure the response time of clinical sta when a fall occurs, and will be a key component in evaluating and improving fall prevention programs. Phase I Results The Phase I e ort built and evaluated the FallCall prototype in both a fall laboratory and real-world hospital setting using patients. The studies resulted in 100% of falls detected with zero false alarms. Location determination also achieved perfect accuracy. A formal usability study was conducted resulting in a grade in the top 5% based on normative data.