Predicting cardiorespiratory instability

Pinsky, Michael R.; Clermont, Gilles; Hravnak, Marilyn

doi:10.1186/s13054-016-1223-7

Cited by 26 publications

(9 citation statements)

References 41 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…For example, support vector machines have been used in noninvasive estimates of simulated and actual hemorrhage severity [100,101], though it should be noted that these models were used to separate 2-3 classes of severity rather than have a continuous output and were also not individual-specific. Other studies have utilized random forest, support vector machines, k-nearest neighbors and neural network algorithms to model risk of cardiorespiratory instability in the hospital through integrated monitoring systems [102][103][104]. K-nearest neighbors, random forest, gradient boosted trees and logistic regression with L2 regularization were used to classify hypotensive events in time series data from the ICU [105].…”

Section: Integrationmentioning

confidence: 99%

Wearable Sensors and Machine Learning for Hypovolemia Problems in Occupational, Military and Sports Medicine: Physiological Basis, Hardware and Algorithms

Kimball

Inan

Convertino

et al. 2022

Sensors

View full text Add to dashboard Cite

Hypovolemia is a physiological state of reduced blood volume that can exist as either (1) absolute hypovolemia because of a lower circulating blood (plasma) volume for a given vascular space (dehydration, hemorrhage) or (2) relative hypovolemia resulting from an expanded vascular space (vasodilation) for a given circulating blood volume (e.g., heat stress, hypoxia, sepsis). This paper examines the physiology of hypovolemia and its association with health and performance problems common to occupational, military and sports medicine. We discuss the maturation of individual-specific compensatory reserve or decompensation measures for future wearable sensor systems to effectively manage these hypovolemia problems. The paper then presents areas of future work to allow such technologies to translate from lab settings to use as decision aids for managing hypovolemia. We envision a future that incorporates elements of the compensatory reserve measure with advances in sensing technology and multiple modalities of cardiovascular sensing, additional contextual measures, and advanced noise reduction algorithms into a fully wearable system, creating a robust and physiologically sound approach to manage physical work, fatigue, safety and health issues associated with hypovolemia for workers, warfighters and athletes in austere conditions.

show abstract

Section: Integrationmentioning

confidence: 99%

Wearable Sensors and Machine Learning for Hypovolemia Problems in Occupational, Military and Sports Medicine: Physiological Basis, Hardware and Algorithms

Kimball

Inan

Convertino

et al. 2022

Sensors

View full text Add to dashboard Cite

show abstract

“…One team from Pittsburgh in the United States has already developed, validated, and tested real-time intraoperative risk prediction tools based on electronic health record data and high-fidelity physiological waveforms to predict cardiopulmonary instability in the intensive care unit. 111 More recently, another team from London developed the "Artificial Intelligence Clinician" to learn optimal treatment strategies for sepsis in intensive care. 112 The authors demonstrated that using their "Artificial Intelligence Clinician" allowed better treatment selection which could lead to lower mortality rates in patients.…”

Section: Future Directionsmentioning

confidence: 99%

Autonomous Systems in Anesthesia: Where Do We Stand in 2020? A Narrative Review

Zaouter

Joosten

Rinehart

et al. 2020

Anesthesia &Amp; Analgesia

View full text Add to dashboard Cite

BIS = bispectral index; CDS = clinical decision support system; EEG = electroencephalogram; HSS = hybrid sedation system; MIMO = multiple input multiple output; SISO = single input single output; TCI = target-controlled infusion BACKGROUND: THE RATIONALE FOR ROBOTS IN ANESTHESIA Robots now surround our daily activities, doing everything from cleaning our homes to flying airplanes, and they exist in fields ranging from industry to medicine.In science fiction, we tend to conceptualize "robots" as human-shaped automata, but in general, a "robot" generally refers to any mechanical system capable of interacting with the environment with directed interventions.In medicine, automation offers precision therapy in combination with a high level of reproducibility. These features make robots appealing in the medical fields, and they have been used in the field of anesthesia for several decades assisting clinicians. 1,2 However, many well-tested autonomous systems and known to be safe have not been adopted yet into clinical practice on a daily basis. When we consider that repetitive execution of trivial technical tasks is subject to vigilance decrement in humans, as well as the fact that human providers may suffer from fatigue, boredom, and bias, there is a strong rationale for the role such systems may play in clinical care. 3,4 Using robots to assist in the acquisition of patient data, simple decision-making, and manual tasks may leave physicians freer to focus efficiently on tasks requiring human intelligence and judgment. 5 As most of us are aware, almost every facet of our society is becoming, for better or worse, progressively more technology-dependent. Technological advancement has made autonomous systems, also known as robots, an integral part of our life in several fields, including medicine. The application of robots in anesthesia could be classified into 3 types of robots. The first ones are pharmacological robots. These robots are based on closed-loop systems that allow betterindividualized anesthetic drug titration for optimal homeostasis during general anesthesia and sedation. Recent evidence also demonstrates that autonomous systems could control hemodynamic parameters proficiently outperforming manual control in the operating room. The second type of robot is mechanical. They enable automated motorized reproduction of tasks requiring high manual dexterity level. Such robots have been advocated to be more accurate than humans and, thus, could be safer for the patient. The third type is a cognitive robot also known as decision support system. This type of robot is able to recognize crucial clinical situation that requires human intervention. When these events occur, the system notifies the attending clinician, describes relevant related clinical observations, proposes pertinent therapeutic options and, when allowed by the attending clinician, may even administer treatment. It seems that cognitive robots could increase patients' safety. Robots in anesthesia offer not only the possibility to free the attending cl...

show abstract

“…A combination of more information sources together (i.e., pulse transit time, finger volume clamp, and surface sensor) may further improve the way we perform hemodynamic monitoring. For instance, taking together more vital signs (like the Vital Sign Index by Visensia™ monitor, OBS Medical, IN-USA) may help to predict cardiac instability ( 51 ) or assessing the heart rate variability from electrocardiography may be useful in predicting hypotension ( 52 ). Another example may be the recently approved Hypotension Probability Indicator by Edwards Lifesciences Inc. which should be able to predict hypotension based on analysis of multiple domains including arterial pressure curve complexity, heart rate variability, and others by proprietary algorithm combined with machine learning.…”

Section: Emerging and Future Conceptsmentioning

confidence: 99%

Continuous Non-Invasive Arterial Pressure Assessment during Surgery to Improve Outcome

Stenglova

Beneš

2017

Front. Med.

View full text Add to dashboard Cite

Blood pressure (BP) is one of the most important variables evaluated during almost every medical examination. Most national anesthesiology societies recommend BP monitoring at least once every 5 min in anesthetized subjects undergoing surgical procedures. In most cases, BP is monitored non-invasively using oscillometric cuffs. Although the risk of arterial cannulation is not very high, the invasive BP monitoring is usually indicated only in the case of high-risk patients or in complex surgical procedures. However, recent evidence points out that when using intermittent BP monitoring short periods of hypotension may be overlooked. In addition, large datasets have demonstrated that even short periods of low BP (or their cumulative duration) may have a detrimental impact on the development of postoperative outcome including increased risk of acute kidney or myocardial injury development. Recently marketed continuous non-invasive blood pressure monitoring tools may help us to recognize the BP fluctuation without the associated burden of arterial cannulation filling the gap between intermittent non-invasive cuff and continuous invasive arterial pressure. Among others, several novel devices based either on volume clamp/vascular unloading method or on applanation tonometry are nowadays available. Moreover, several near-future smart technologies may lead to better hypotension recognition or even prediction potentially improving our ability to maintain BP stability throughout the anesthesia or surgical procedure. In this review, novel or emerging technologies of non-invasive continuous blood pressure assessment and their potential to improve postoperative outcome are discussed.

show abstract

Predicting cardiorespiratory instability

Cited by 26 publications

References 41 publications

Wearable Sensors and Machine Learning for Hypovolemia Problems in Occupational, Military and Sports Medicine: Physiological Basis, Hardware and Algorithms

Wearable Sensors and Machine Learning for Hypovolemia Problems in Occupational, Military and Sports Medicine: Physiological Basis, Hardware and Algorithms

Autonomous Systems in Anesthesia: Where Do We Stand in 2020? A Narrative Review

Continuous Non-Invasive Arterial Pressure Assessment during Surgery to Improve Outcome

Contact Info

Product

Resources

About