Improving monitoring data utility in acute care medicine using mechanistic mathematical models of physiological processes

在过去的几十年里，医疗监测技术有了很大的进步。在麻醉学和重症监护的数据特别丰富的环境中，这些进步在很大程度上未能转化为改善的结果。这可能至少部分归因于人类无法完全吸收、定量解释和利用生物信号及其相互作用的多模态时间序列中包含的定量信息进行患者状态评估、预测和治疗优化，这表明计算机化决策支持的作用。本提案中概述的研究计划将测试以下假设：临床数据解释、预测和治疗优化基于定量描述高度非线性的生理学机械数学模型的概率反演，产生观察结果的耦合生理过程可以帮助缓解这种情况，并为麻醉学和重症监护中的医生提供计算机化的决策支持工具，从所有可用的监测模式的临床数据，并提供输出，是很容易获得的生理和临床解释，有可能实现一个完全个性化的方法，急性护理，但仍然是访问验证满足循证医学的标准。拟议项目将侧重于纯观察性研究中的模型和方法开发与验证，旨在评估和预测血流动力学状态，为将来对所开发方法的临床有用性进行干预性评价奠定基础。

Medical monitoring technology has advanced significantly over the past decades. In the particularly data rich environments of anaesthesiology and intensive care, these advancements have largely failed to translate into improved outcomes. This may at least partially be attributable to the human inability to fully assimilate, quantitatively interpret, and utilize for patient status assessment, prediction, and therapy optimization the quantitative information contained in the multimodal time-series of biosignals and their interactions, suggesting a role for computerized decision support. The research plan outlined in this proposal will test the hypothesis that clinical data interpretation, prediction, and therapeutic optimization based on probabilistic inversion of mechanistic mathematical models of physiology describing quantitatively the highly nonlinear, coupled physiological processes generating the observations can help to alleviate this situation and provide physicians in anaesthesiology and intensive care with a computerized decision support tool that integrates clinical data from all available monitoring modalities and provides output that is readily accessible to physiological and clinical interpretation, with the potential to realize a fully personalized approach to acute care that is nevertheless accessible to validation satisfying the criteria of Evidence Based Medicine. The proposed project will focus on model and methods development and validation in purely observational studies targeting assessment and prediction of hemodynamic status to set the stage for future interventional evaluations of the clinical usefulness of the developed methodologies.