Patient-Specific Classification of ICU Sedation Levels From Heart Rate Variability

A deep learning model for estimating sedation levels using heart rate variability and vital signs: a retrospective cross-sectional study at a center in South Korea

BACKGROUND

Optimal sedation assessment in critically ill children remains challenging due to the subjective nature of behavioral scales and intermittent evaluation schedules. This study aimed to develop a deep learning model based on heart rate variability (HRV) parameters and vital signs to predict effective and safe sedation levels in pediatric patients.

METHODS

This retrospective cross-sectional study was conducted in a pediatric intensive care unit at a tertiary children's hospital. We developed deep learning models incorporating HRV parameters extracted from electrocardiogram waveforms and vital signs to predict Richmond Agitation-Sedation Scale (RASS) scores. Model performance was evaluated using the area under the receiver operating characteristic curve (AUROC) and area under the precision-recall curve (AUPRC). The data were split into training, validation, and test sets (6:2:2), and the models were developed using a 1D ResNet architecture.

RESULTS

Analysis of 4,193 feature sets from 324 patients achieved excellent discrimination ability, with AUROC values of 0.867, 0.868, 0.858, 0.851, and 0.811 for whole number RASS thresholds of -5 to -1, respectively. AUPRC values ranged from 0.928 to 0.623, showing superior performance in deeper sedation levels. The HRV metric SDANN2 showed the highest feature importance, followed by systolic blood pressure and heart rate.

CONCLUSIONS

A combination of HRV parameters and vital signs can effectively predict sedation levels in pediatric patients, offering the potential for automated and continuous sedation monitoring in pediatric intensive care settings. Future multi-center validation studies are needed to establish broader applicability.

Establishment and Verification of an Artificial Intelligence Prediction Model for Children With Sepsis

BACKGROUND

Early identification of high-risk groups of children with sepsis is beneficial to reduce sepsis mortality. This article used artificial intelligence (AI) technology to predict the risk of death effectively and quickly in children with sepsis in the pediatric intensive care unit (PICU).

STUDY DESIGN

This retrospective observational study was conducted in the PICUs of the First Affiliated Hospital of Sun Yat-sen University from December 2016 to June 2019 and Shenzhen Children's Hospital from January 2019 to July 2020. The children were divided into a death group and a survival group. Different machine language (ML) models were used to predict the risk of death in children with sepsis.

RESULTS

A total of 671 children with sepsis were enrolled. The accuracy (ACC) of the artificial neural network model was better than that of support vector machine, logical regression analysis, Bayesian, K nearest neighbor method and decision tree models, with a training set ACC of 0.99 and a test set ACC of 0.96.

CONCLUSIONS

The AI model can be used to predict the risk of death due to sepsis in children in the PICU, and the artificial neural network model is better than other AI models in predicting mortality risk.

Artificial Intelligence and Machine Learning Applications in Critically Ill Brain Injured Patients

The utilization of Artificial Intelligence (AI) and Machine Learning (ML) is paving the way for significant strides in patient diagnosis, treatment, and prognostication in neurocritical care. These technologies offer the potential to unravel complex patterns within vast datasets ranging from vast clinical data and EEG (electroencephalogram) readings to advanced cerebral imaging facilitating a more nuanced understanding of patient conditions. Despite their promise, the implementation of AI and ML faces substantial hurdles. Historical biases within training data, the challenge of interpreting multifaceted data streams, and the "black box" nature of ML algorithms present barriers to widespread clinical adoption. Moreover, ethical considerations around data privacy and the need for transparent, explainable models remain paramount to ensure trust and efficacy in clinical decision-making.This article reflects on the emergence of AI and ML as integral tools in neurocritical care, discussing their roles from the perspective of both their scientific promise and the associated challenges. We underscore the importance of extensive validation in diverse clinical settings to ensure the generalizability of ML models, particularly considering their potential to inform critical medical decisions such as withdrawal of life-sustaining therapies. Advancement in computational capabilities is essential for implementing ML in clinical settings, allowing for real-time analysis and decision support at the point of care. As AI and ML are poised to become commonplace in clinical practice, it is incumbent upon health care professionals to understand and oversee these technologies, ensuring they adhere to the highest safety standards and contribute to the realization of personalized medicine. This engagement will be pivotal in integrating AI and ML into patient care, optimizing outcomes in neurocritical care through informed and data-driven decision-making.

Artificial Intelligence in Perioperative Planning and Management of Liver Resection

Artificial intelligence (AI) is a speciality within computer science that deals with creating systems that can replicate the intelligence of a human mind and has problem-solving abilities. AI includes a diverse array of techniques and approaches such as machine learning, neural networks, natural language processing, robotics, and expert systems. An electronic literature search was conducted using the databases of "PubMed" and "Google Scholar". The period for the search was from 2000 to June 2023. The search terms included "artificial intelligence", "machine learning", "liver cancers", "liver tumors", "hepatectomy", "perioperative" and their synonyms in various combinations. The search also included all MeSH terms. The extracted articles were further reviewed in a step-wise manner for identification of relevant studies. A total of 148 articles were identified after the initial literature search. Initial review included screening of article titles for relevance and identifying duplicates. Finally, 65 articles were reviewed for this review article. The future of AI in liver cancer planning and management holds immense promise. AI-driven advancements will increasingly enable precise tumour detection, location, and characterisation through enhanced image analysis. ML algorithms will predict patient-specific treatment responses and complications, allowing for tailored therapies. Surgical robots and AI-guided procedures will enhance the precision of liver resections, reducing risks and improving outcomes. AI will also streamline patient monitoring, better hemodynamic management, enabling early detection of recurrence or complications. Moreover, AI will facilitate data-driven research, accelerating the development of novel treatments and therapies. Ultimately, AI's integration will revolutionise liver cancer care, offering personalised, efficient and effective solutions, improving patients' quality of life and survival rates.

Role of artificial intelligence in perioperative monitoring in anaesthesia

Artificial intelligence (AI) is making giant strides in the medical domain, and the field of anaesthesia is not untouched. Enhancement in technology, especially AI, in many fields, including medicine, has proven to be far superior, safer and less erratic than human decision-making. The intersection of anaesthesia and AI holds the potential for augmenting constructive advances in anaesthesia care. AI can improve anaesthesiologists' efficiency, reduce costs and improve patient outcomes. Anaesthesiologists are well placed to harness the advantages of AI in various areas like perioperative monitoring, anaesthesia care, drug delivery, post-anaesthesia care unit, pain management and intensive care unit. Perioperative monitoring of the depth of anaesthesia, clinical decision support systems and closed-loop anaesthesia delivery aid in efficient and safer anaesthesia delivery. The effect of various AI interventions in clinical practice will need further research and validation, as well as the ethical implications of privacy and data handling. This paper aims to provide an overview of AI in perioperative monitoring in anaesthesia.

Classification of Level of Consciousness in a Neurological ICU Using Physiological Data

BACKGROUND

Impaired consciousness is common in intensive care unit (ICU) patients, and an individual's degree of consciousness is crucial to determining their care and prognosis. However, there are no methods that continuously monitor consciousness and alert clinicians to changes. We investigated the use of physiological signals collected in the ICU to classify levels of consciousness in critically ill patients.

METHODS

We studied 61 patients with subarachnoid hemorrhage (SAH) and 178 patients with intracerebral hemorrhage (ICH) from the neurological ICU at Columbia University Medical Center in a retrospective observational study of prospectively collected data. The level of consciousness was determined on the basis of neurological examination and mapped to comatose, vegetative state or unresponsive wakefulness syndrome (VS/UWS), minimally conscious minus state (MCS-), and command following. For each physiological signal, we extracted time-series features and performed classification using extreme gradient boosting on multiple clinically relevant tasks across subsets of physiological signals. We applied this approach independently on both SAH and ICH patient groups for three sets of variables: (1) a minimal set common to most hospital patients (e.g., heart rate), (2) variables available in most ICUs (e.g., body temperature), and (3) an extended set recorded mainly in neurological ICUs (absent for the ICH patient group; e.g., brain temperature).

RESULTS

On the commonly performed classification task of VS/UWS versus MCS-, we achieved an area under the receiver operating characteristic curve (AUROC) in the SAH patient group of 0.72 (sensitivity 82%, specificity 57%; 95% confidence interval [CI] 0.63-0.81) using the extended set, 0.69 (sensitivity 83%, specificity 51%; 95% CI 0.59-0.78) on the variable set available in most ICUs, and 0.69 (sensitivity 56%, specificity 78%; 95% CI 0.60-0.78) on the minimal set. In the ICH patient group, AUROC was 0.64 (sensitivity 56%, specificity 65%; 95% CI 0.55-0.74) using the minimal set and 0.61 (sensitivity 50%, specificity 80%; 95% CI 0.51-0.71) using the variables available in most ICUs.

CONCLUSIONS

We find that physiological signals can be used to classify states of consciousness for patients in the ICU. Building on this with intraday assessments and increasing sensitivity and specificity may enable alarm systems that alert physicians to changes in consciousness and frequent monitoring of consciousness throughout the day, both of which may improve patient care and outcomes.

Early heart rate variability evaluation enables to predict ICU patients' outcome

Heart rate variability (HRV) is a mean to evaluate cardiac effects of autonomic nervous system activity, and a relation between HRV and outcome has been proposed in various types of patients. We attempted to evaluate the best determinants of such variation in survival prediction using a physiological data-warehousing program. Plethysmogram tracings (PPG) were recorded at 75 Hz from the standard monitoring system, for a 2 h period, during the 24 h following ICU admission. Physiological data recording was associated with metadata collection. HRV was derived from PPG in either the temporal and non-linear domains. 540 consecutive patients were recorded. A lower LF/HF, SD2/SD1 ratios and Shannon entropy values on admission were associated with a higher ICU mortality. SpO2/FiO2 ratio and HRV parameters (LF/HF and Shannon entropy) were independent correlated with mortality in the multivariate analysis. Machine-learning using neural network (kNN) enabled to determine a simple decision tree combining the three best determinants (SDNN, Shannon Entropy, SD2/SD1 ratio) of a composite outcome index. HRV measured on admission enables to predict outcome in the ICU or at Day-28, independently of the admission diagnosis, treatment and mechanical ventilation requirement.

Trial registration: ClinicalTrials.gov identifier NCT02893462.

Risk factors, time to onset and recurrence of delirium in a mixed medical-surgical ICU population: A secondary analysis using Cox and CHAID decision tree modeling

A retrospective secondary analysis of 4,200 patients was collected from two academic medical centers. Delirium was assessed using the Confusion Assessment Method for the Intensive Care Unit (CAM-ICU) in all patients. Univariate and multivariate Cox models, logistic regression analysis, and Chi-square Automatic Interaction Detector (CHAID) decision tree modeling were used to explore delirium risk factors. Increased delirium risk was associated with exposed only to artificial light (AL) hazard ratio (HR) 1.84 (95 % CI: 1.66-2.044, P<0.001), physical restraint application 1.11 (95 % CI: 1.001-1.226, P=0.049), and high nursing care requirements (>8 hours per 8-hour shift) 1.18 (95 % CI: 1.048-1.338, P=0.007). Delirium incidence was inversely associated with greater family engagement 0.092 (95 % CI: 0.014-0.596, P=0.012), low staff burnout and anticipated turnover scores 0.093 (95 % CI: 0.014-0.600, P=0.013), non-ICU length-of-stay (LOS)<15 days 0.725 (95 % CI: 0.655-0.804, P<0.001), and ICU LOS ≤15 days 0.509 (95 % CI: 0.456-0.567, P<0.001). CHAID modeling indicated that AL exposure and age <65 years were associated with a high risk of delirium incidence, whereas SOFA score ≤11, APACHE IV score >15 and natural light (NL) exposure were associated with moderate risk, and female sex was associated with low risk. More rapid time to delirium onset correlated with baseline sleep disturbance (P=0.049), high nursing care requirements (P=0.019), and prolonged ICU and non-ICU hospital LOS (P<0.001). Delirium recurrence correlated with age >65 years (HR 2.198; 95 % CI: 1.101-4.388, P=0.026) and high nursing care requirements (HR 1.978, 95 % CI: 1.096-3.569), with CHAID modeling identifying AL exposure (P<0.001) and age >65 years (P=0.032) as predictive variables. Development of ICU delirium correlated with application of physical restraints, high nursing care requirements, prolonged ICU and non-ICU LOS, exposure exclusively to AL (rather than natural), less family engagement, and greater staff burnout and anticipated turnover scores. ICU delirium occurred more rapidly in patients with baseline sleep disturbance, and recurrence correlated with the presence of delirium on ICU admission, exclusive AL exposure, and high nursing care requirements.

Machine Learning, Deep Learning, and Closed Loop Devices-Anesthesia Delivery

With the tremendous volume of data captured during surgeries and procedures, critical care, and pain management, the field of anesthesiology is uniquely suited for the application of machine learning, neural networks, and closed loop technologies. In the past several years, this area has expanded immensely in both interest and clinical applications. This article provides an overview of the basic tenets of machine learning, neural networks, and closed loop devices, with emphasis on the clinical applications of these technologies.

Impact of Religiosity on Delirium Severity Among Critically Ill Shi'a Muslims: A Prospective Multi-Center Observational Study

This study assesses the impact of religiosity on delirium severity and patient outcomes among Shi'a Muslim intensive care unit (ICU) patients. We conducted a prospective observational cohort study in 21 ICUs from 6 Iranian academic medical centers. Delirium was assessed using the Confusion Assessment Method for the ICU (CAM-ICU) tool. Eligible patients were intubated, receiving mechanical ventilation (MV) for ≥ 48 h. Illness severity was assessed using Acute Physiology and Chronic Health Evaluation II (APACHE II) scores. A total of 4200 patients were enrolled. Patient religiosity was categorized as more (40.6%), moderate (42.3%), or less (17.1%) based on responses to patient and surrogate questionnaires. The findings suggest that lower pre-illness religiosity may be associated with greater delirium severity, MV duration, and ICU and hospital LOS. The lower mortality in the less religiosity group may be related in part to a greater proportion of female patients, but it remains unclear whether and to what extent greater religiosity impacted treatment decisions by patients and families. Further investigation is needed to validate and clarify the mechanism of the mortality findings.

Adding Continuous Vital Sign Information to Static Clinical Data Improves the Prediction of Length of Stay After Intubation: A Data-Driven Machine Learning Approach

BACKGROUND

Bedside monitors in the ICU routinely measure and collect patients' physiologic data in real time to continuously assess the health status of patients who are critically ill. With the advent of increased computational power and the ability to store and rapidly process big data sets in recent years, these physiologic data show promise in identifying specific outcomes and/or events during patients' ICU hospitalization.

METHODS

We introduced a methodology designed to automatically extract information from continuous-in-time vital sign data collected from bedside monitors to predict if a patient will experience a prolonged stay (length of stay) on mechanical ventilation, defined as >4 d, in a pediatric ICU.

RESULTS

Continuous-in-time vital signs information and clinical history data were retrospectively collected for 284 ICU subjects from their first 24 h on mechanical ventilation from a medical-surgical pediatric ICU at Boston Children's Hospital. Multiple machine learning models were trained on multiple subsets of these subjects to predict the likelihood that each of these subjects would experience a long stay. We evaluated the predictive power of our models strictly on unseen hold-out validation sets of subjects. Our methodology achieved model performance of >83% (area under the curve) by using only vital sign information as input, and performances of 90% (area under the curve) by combining vital sign information with subjects' static clinical data readily available in electronic health records. We implemented this approach on 300 independently trained experiments with different choices of training and hold-out validation sets to ensure the consistency and robustness of our results in our study sample. The predictive power of our approach outperformed recent efforts that used deep learning to predict a similar task.

CONCLUSIONS

Our proposed workflow may prove useful in the design of scalable approaches for real-time predictive systems in ICU environments, exploiting real-time vital sign information from bedside monitors. (ClinicalTrials.gov registration NCT02184208.).

Predicting Acute Kidney Injury after Cardiac Surgery by Machine Learning Approaches

Cardiac surgery-associated AKI (CSA-AKI) is common after cardiac surgery and has an adverse impact on short- and long-term mortality. Early identification of patients at high risk of CSA-AKI by applying risk prediction models allows clinicians to closely monitor these patients and initiate effective preventive and therapeutic approaches to lessen the incidence of AKI. Several risk prediction models and risk assessment scores have been developed for CSA-AKI. However, the definition of AKI and the variables utilized in these risk scores differ, making general utility complex. Recently, the utility of artificial intelligence coupled with machine learning, has generated much interest and many studies in clinical medicine, including CSA-AKI. In this article, we discussed the evolution of models established by machine learning approaches to predict CSA-AKI.

Artificial Intelligence in Anesthesiology: Current Techniques, Clinical Applications, and Limitations

Artificial intelligence has been advancing in fields including anesthesiology. This scoping review of the intersection of artificial intelligence and anesthesia research identified and summarized six themes of applications of artificial intelligence in anesthesiology: (1) depth of anesthesia monitoring, (2) control of anesthesia, (3) event and risk prediction, (4) ultrasound guidance, (5) pain management, and (6) operating room logistics. Based on papers identified in the review, several topics within artificial intelligence were described and summarized: (1) machine learning (including supervised, unsupervised, and reinforcement learning), (2) techniques in artificial intelligence (e.g., classical machine learning, neural networks and deep learning, Bayesian methods), and (3) major applied fields in artificial intelligence.The implications of artificial intelligence for the practicing anesthesiologist are discussed as are its limitations and the role of clinicians in further developing artificial intelligence for use in clinical care. Artificial intelligence has the potential to impact the practice of anesthesiology in aspects ranging from perioperative support to critical care delivery to outpatient pain management.

Automatic Detection of General Anesthetic-States using ECG-Derived Autonomic Nervous System Features

Electroencephalogram (EEG)-based prediction systems are used to target anesthetic-states in patients undergoing procedures with general anesthesia (GA). These systems are not widely employed in resource-limited settings because they are cost-prohibitive. Although anesthetic-drugs induce highly-structured, oscillatory neural dynamics that make EEG-based systems a principled approach for anesthetic-state monitoring, anesthetic-drugs also significantly modulate the autonomic nervous system (ANS). Because ANS dynamics can be inferred from electrocardiogram (ECG) features such as heart rate variability, it may be possible to develop an ECG-based system to infer anesthetic-states as a low-cost and practical alternative to EEG-based anesthetic-state prediction systems. In this work, we demonstrate that an ECG-based system using ANS features can be used to discriminate between non-GA and GA states in sevoflurane, with a GA F₁ score of 0.834, [95% CI, 0.776, 0.892], and in sevoflurane-plus-ketamine, with a GA F₁ score of 0.880 [0.815, 0.954]. With further refinement, ECG-based anesthetic-state systems could be developed as a fully automated system for anesthetic-state monitoring in resource-limited settings.

Artificial Intelligence and Machine Learning in Anesthesiology

Commercial applications of artificial intelligence and machine learning have made remarkable progress recently, particularly in areas such as image recognition, natural speech processing, language translation, textual analysis, and self-learning. Progress had historically languished in these areas, such that these skills had come to seem ineffably bound to intelligence. However, these commercial advances have performed best at single-task applications in which imperfect outputs and occasional frank errors can be tolerated.The practice of anesthesiology is different. It embodies a requirement for high reliability, and a pressured cycle of interpretation, physical action, and response rather than any single cognitive act. This review covers the basics of what is meant by artificial intelligence and machine learning for the practicing anesthesiologist, describing how decision-making behaviors can emerge from simple equations. Relevant clinical questions are introduced to illustrate how machine learning might help solve them-perhaps bringing anesthesiology into an era of machine-assisted discovery.

Brain Monitoring of Sedation in the Intensive Care Unit Using a Recurrent Neural Network

Over and under-sedation are common in critically ill patients admitted to the Intensive Care Unit. Clinical assessments provide limited time resolution and are based on behavior rather than the brain itself. Existing brain monitors have been developed primarily for non-ICU settings. Here, we use a clinical dataset from 154 ICU patients in whom the Richmond Agitation-Sedation Score is assessed about every 2 hours. We develop a recurrent neural network (RNN) model to discriminate between deep vs. no sedation, trained end-to-end from raw EEG spectrograms without any feature extraction. We obtain an average area under the ROC of 0.8 on 10-fold cross validation across patients. Our RNN is able to provide reliable estimates of sedation levels consistently better compared to a feed-forward model with simple smoothing. Decomposing the prediction error in terms of sedatives reveals that patient-specific calibration for sedatives is expected to further improve sedation monitoring.

Intensive Care Unit Telemedicine in the Era of Big Data, Artificial Intelligence, and Computer Clinical Decision Support Systems

This article examines the history of the telemedicine intensive care unit (tele-ICU), the current state of clinical decision support systems (CDSS) in the tele-ICU, applications of machine learning (ML) algorithms to critical care, and opportunities to integrate ML with tele-ICU CDSS. The enormous quantities of data generated by tele-ICU systems is a major driver in the development of the large, comprehensive, heterogeneous, and granular data sets necessary to develop generalizable ML CDSS algorithms, and deidentification of these data sets expands opportunities for ML CDSS research.

Potential EEG biomarkers of sedation doses in intensive care patients unveiled by using a machine learning approach

OBJECTIVE

Sedation of neurocritically ill patients is one of the most challenging situation in ICUs. Quantitative knowledge on the sedation effect on brain activity in that complex scenario could help to uncover new markers for sedation assessment. Hence, we aim to evaluate the existence of changes of diverse EEG-derived measures in deeply-sedated (RASS-Richmond agitation-sedation scale  -4 and  -5) neurocritically ill patients, and also whether sedation doses are related with those eventual changes.

APPROACH

We performed an observational prospective cohort study in the intensive care unit of the Hospital de la Princesa. Twenty-six adult patients suffered from traumatic brain injury and subarachnoid hemorrhage were included in the present study. Long-term continuous electroencephalographic (EEG) recordings (2141 h) and hourly annotated information were used to determine the relationship between intravenous sedation infusion doses and network and spectral EEG measures. To do that, two different strategies were followed: assessment of the statistical dependence between both variables using the Spearman correlation rank and by performing an automatic classification method based on a machine learning algorithm.

MAIN RESULTS

More than 60% of patients presented a correlation greater than 0.5 in at least one of the calculated EEG measures with the sedation dose. The automatic classification method presented an accuracy of 84.3% in discriminating between different sedation doses. In both cases the nodes' degree was the most relevant measurement.

SIGNIFICANCE

The results presented here provide evidences of brain activity changes during deep sedation linked to sedation doses. Particularly, the capability of network EEG-derived measures in discriminating between different sedation doses could be the framework for the development of accurate methods for sedation levels assessment.

Comparison of non-invasive to invasive oxygenation ratios for diagnosing acute respiratory distress syndrome following coronary artery bypass graft surgery: a prospective derivation-validation cohort study

Objective

To determine if non-invasive oxygenation indices, namely peripheral capillary oxygen saturation (S_pO₂)/fraction of inspired oxygen (F_iO₂) and partial pressure of alveolar oxygen (P_AO₂)/F_iO₂ may be used as effective surrogates for the partial pressure of arterial oxygen (P_aO₂)/F_iO₂. Also, to determine the S_pO₂/F_iO₂ and P_AO₂/F_iO₂ values that correspond to P_aO₂/F_iO₂ thresholds for identifying acute respiratory distress syndrome (ARDS) in patients following coronary artery bypass graft (CABG) surgery.

Methods

A prospective derivation-validation cohort study in the Open-Heart ICU of an academic teaching hospital. Recorded variables included patient demographics, ventilator settings, chest radiograph results, and S_PO₂, P_aO₂, P_AO_2, S_aO₂, and F_iO₂. Linear regression modeling was used to quantify the relationship between indices. Receiver operating characteristic (ROC) curves were used to determine the sensitivity and specificity of the threshold values.

Results

One-hundred seventy-five patients were enrolled in the derivation cohort, and 358 in the validation cohort. The S_PO₂/F_iO₂ and P_AO₂/F_iO₂ ratios could be predicted well from P_aO₂/F_iO₂, described by the linear regression models S_PO₂/F_iO₂ = 71.149 + 0.8PF and P_AO₂/F_iO₂ = 38.098 + 2.312PF, respectively. According to the linear regression equation, a P_aO₂/F_iO₂ ratio of 300 equaled an S_PO₂/F_iO₂ ratio of 311 (R² 0.857, F 1035.742, < 0.0001) and a P_AO₂/F_iO₂ ratio of 732 (R² 0.576, F 234.887, < 0.0001). The S_PO₂/F_iO₂ threshold of 311 had 90% sensitivity, 80% specificity, LR+ 4.50, LR- 0.13, PPV 98, and NPV 42.1 for the diagnosis of mild ARDS. The P_AO₂/F_iO₂ threshold of 732 had 86% sensitivity, 90% specificity, LR+ 8.45, LR- 0.16, PPV 98.9, and NPV 36 for the diagnosis of mild ARDS. S_PO₂/F_iO₂ had excellent discrimination ability for mild ARDS (AUC ± SE = 0.92 ± 0.017; 95% CI 0.889 to 0.947) as did P_AO₂/F_iO₂ (AUC ± SE = 0.915 ± 0.018; 95% CI 0.881 to0.942).

Conclusions

P_aO₂ and S_aO₂ correlated in the diagnosis of ARDS, with a P_aO₂/F_iO₂ of 300 correlating to an S_PO₂/ F_iO₂ of 311 (Sensitivity 90%, Specificity 80%). The S_PO₂/ F_iO₂ ratio may allow for early real-time rapid identification of ARDS, while decreasing the cost, phlebotomy, blood loss, pain, skin breaks, and vascular punctures associated with serial arterial blood gas measurements.

Applying machine learning to continuously monitored physiological data

The use of machine learning (ML) in healthcare has enormous potential for improving disease detection, clinical decision support, and workflow efficiencies. In this commentary, we review published and potential applications for the use of ML for monitoring within the hospital environment. We present use cases as well as several questions regarding the application of ML to the analysis of the vast amount of complex data that clinicians must interpret in the realm of continuous physiological monitoring. ML, especially employed in bidirectional conjunction with electronic health record data, has the potential to extract much more useful information out of this currently under-analyzed data source from a population level. As a data driven entity, ML is dependent on copious, high quality input data so that error can be introduced by low quality data sources. At present, while ML is being studied in hybrid formulations along with static expert systems for monitoring applications, it is not yet actively incorporated in the formal artificial learning sense of an algorithm constantly learning and updating its rules without external intervention. Finally, innovations in monitoring, including those supported by ML, will pose regulatory and medico-legal challenges, as well as questions regarding precisely how to incorporate these features into clinical care and medical education. Rigorous evaluation of ML techniques compared to traditional methods or other AI methods will be required to validate the algorithms developed with consideration of database limitations and potential learning errors. Demonstration of value on processes and outcomes will be necessary to support the use of ML as a feature in monitoring system development: Future research is needed to evaluate all AI based programs before clinical implementation in non-research settings.