The Detection, Quantification, and Management of Ventilator Dyssynchrony

项目摘要/摘要 尽管在呼吸机管理方面取得了重大进展，但急性呼吸窘迫综合征的死亡率仍在下降。 Drome(ARDS)仍然高得令人无法接受。大潮气量、高压机械通风 倾斜，反复肺泡塌陷会损伤肺，称为呼吸机诱导的肺损伤(VILI)。已定义 呼吸机不同步(VD)是由于患者努力呼吸的时机和时机不当。 可能会增强Vili的作用。这份提案概述了一项为期5年的培训计划，包括指导、正式教学、 和实际研究经验，这些经验为索蒂尔博士成功的临床研究生涯检查奠定了基础 呼吸机不同步(VD)及其对呼吸机所致肺损伤(VILI)的影响及其优化处理。 综合课程将优化自动VD检测算法，描述哪些类型的VD 是有害的，并确定理想的呼吸机和镇静策略，以最大限度地减少VD并改善 病人的结果。这一经验将为索蒂尔博士提供必要的工具，使其成为信号领域的领导者 分析、呼吸机不同步、高级数学建模和重症监护研究。这一计划将 由机器学习、机械通风、重症监护方面的四位知名专家提供正式指导 动态系统的研究和建模。此外，临床科学、机器学习和 先进的数学模型将为应用这些技术奠定理论基础。这是结构化的 课程将帮助索蒂尔博士从实时呼吸机数据中获得计算机检测VD的专业知识， 血管性痴呆的病理生理学，以及复杂的、时间动态的生物系统的建模。 正式课程将与三次实践经验相吻合。首先，索蒂尔博士将优化他已经 开发了VD识别算法，以检测可能对肺部造成损害的其他类型的VD。第二, 他将确定哪些类型的VD与有害的呼吸机机械有关。最后，他会阻止- 挖掘个性化呼吸机和镇静策略以最大限度地减少VD、VILI和 使用线性最优化和滞后线性相关来考虑患者的动态性质的过度镇静 生理学。 建议的研究结果将开发一种计算机化的算法来检测七种类型的VD， 确定哪些类型的VD最有可能传播VILI，并确定最佳呼吸机和 改善个体患者预后的镇静策略。这将留给索蒂尔博士和研究团队 准备进行一项随机临床试验，比较计算机预测的个体化呼吸机和Se-Se呼吸机- 与目前的护理标准相比较的筹资策略。如果成功，这可能会给 通过为个体患者量身定做循证治疗来使用机械通气。

Project Summary/Abstract Despite significant advances in ventilator management, mortality from the acute respiratory distress syn- drome (ARDS) remains unacceptably high. Mechanical ventilation with large tidal volume, high pressure ven- tilation, and repeated alveolar collapse can injure the lung, called ventilator induced lung injury (VILI). Defined as the inappropriate timing and delivery of a breath in response to a patient effort, ventilator dyssynchrony (VD) may potentiate VILI. This proposal outlines a 5-year training programing including mentoring, formal didactics, and practical research experiences which positions Dr. Sottile for a successful clinical research career examining ventilator dyssynchrony (VD), its impact on ventilator induced lung injury (VILI), and its optimal management. An integrated curriculum will optimize an automated VD detection algorithm, delineate which types of VD are deleterious, and determine the ideal ventilator and sedation strategies to minimize VD and improve patient outcomes. This experience will provide Dr. Sottile with the necessary tools to be a leader in signal analysis, ventilator dyssynchrony, advanced mathematical modeling, and critical care research. This program will consist of formal mentoring from four renowned experts in machine learning, mechanical ventilation, critical care research, and modeling of dynamic systems. In addition, coursework in clinical sciences, machine learning, and advanced mathematical modeling will build the theoretical foundation to apply these techniques. This structured curriculum will help Dr. Sottile gain expertise in the computerized detection of VD from real-time ventilator data, the pathophysiology of VD, and the modeling of complex, temporally dynamic, biological systems. The formal curriculum will coincide with three practical experiences. First, Dr. Sottile will optimize his already developed VD identification algorithm to detect additional types of VD that may be injurious to the lung. Second, he will identify which types of VD are associated with deleterious ventilator mechanics. Finally, he will deter- mine the personalized ventilator and sedation strategies to minimize VD, VILI, and the negative consequences of over sedation using linear optimization and lagged linear correlation to account for the dynamic nature of patient physiology. The result of the proposed studies will develop a computerized algorithm to detect seven types of VD, identify which types of VD are most likely to propagate VILI, and determine the optimal ventilator and sedation strategies to improve individual patient outcomes. This will leave Dr. Sottile and the study team positioned to conduct a randomized clinical trial comparing computer-predicted individualized ventilator and se- dation strategies compared to the current standard of care. If successful, this will potentially to revolutionize the use of mechanical ventilation by tailoring evidenced-based therapies to the individual patient.