Predictive Monitoring: IMPact of Real-time Predictive Monitoring in Acute Care Cardiology Trial (PM-IMPACCT)

点击翻译按钮获取中文摘要

Keim-Malpass/Bourque PM-IMPACCT PROJECT SUMMARY/ABSTRACT Patients on the acute care wards who deteriorate and are emergently transferred to the intensive care unit have poor outcomes. Early identification of subtly worsening patients might allow for earlier clinical action leading to reduced morbidity and mortality. Advances in bedside continuous predictive analytics monitoring (i.e. artificial intelligence (AI)-based risk prediction) make a wealth of data available to healthcare providers and can form the foundation for computational algorithms that integrate real-time bedside monitor physiologic data to provide early warning of potentially catastrophic physiologic events. The future of acute hospital care includes monitoring systems that integrate data streams of rapidly changing clinical information to estimate and communicate risk of imminent events. This will allow a paradigm change in care from reactive to proactive by predicting patient trajectories and acting early to promote optimal patient trajectories. Predictive analytics monitoring is a promising technology that will yield families of these new tools. Here, we propose a multi- disciplinary cluster randomized controlled trial (NCT04359641) to test the use of CoMET (Continuous Monitoring of Event Trajectories), an AI-based visual analytic that displays risk estimates for multiple adverse outcomes. It is expected that having access to a visual risk analytic for impending catastrophic outcomes can draw the clinician’s attention to patients warranting early or extra consideration. Specifically, in our proposed cluster RCT we will evaluate the impact of predictive analytics monitoring on: (1) improvement in patient outcomes, (2) response time to proactive clinical action, and (3) costs to the healthcare system. This proposal is led by an immensely promising interdisciplinary mPI early stage investigators who are members of the Center for Advanced Medical Analytics. This proposal will strengthen the ability of health care organizations to evolve as learning health systems that apply bioinformatics data to improve patient outcomes by incorporating artificial intelligence into knowledge tools that are successfully integrated for use by health care providers and determine if they improve patient outcomes. We anticipate developing standard processes that can be leveraged and are scalable for general implementation of predictive analytics monitoring algorithms in real-life practice contexts. Additionally, we anticipate building on this R01 with future work including a multi-center randomized control trial testing effectiveness of our artificial intelligence-based risk analytic.