A Machine Learning Algorithm to Predict Severe Sepsis and Septic Shock: Development, Implementation, and Impact on Clinical Practice. Crit Care Med. 2019;47:1485-1492. [PMID: 31389839 DOI: 10.1097/ccm.0000000000003891]

Enhancing Patient Selection in Sepsis Clinical Trials Design Through an AI Enrichment Strategy: Algorithm Development and Validation

BACKGROUND

Sepsis is a heterogeneous syndrome, and enrollment of more homogeneous patients is essential to improve the efficiency of clinical trials. Artificial intelligence (AI) has facilitated the identification of homogeneous subgroups, but how to estimate the uncertainty of the model outputs when applying AI to clinical decision-making remains unknown.

OBJECTIVE

We aimed to design an AI-based model for purposeful patient enrollment, ensuring that a patient with sepsis recruited into a trial would still be persistently ill by the time the proposed therapy could impact patient outcome. We also expected that the model could provide interpretable factors and estimate the uncertainty of the model outputs at a customized confidence level.

METHODS

In this retrospective study, 9135 patients with sepsis requiring vasopressor treatment within 24 hours after sepsis onset were enrolled from Beth Israel Deaconess Medical Center. This cohort was used for model development, and 10-fold cross-validation with 50 repeats was used for internal validation. In total, 3743 patients with sepsis from the eICU Collaborative Research Database were used as the external validation cohort. All included patients with sepsis were stratified based on disease progression trajectories: rapid death, recovery, and persistent ill. A total of 148 variables were selected for predicting the 3 trajectories. Four machine learning algorithms with 3 different setups were used. We estimated the uncertainty of the model outputs using conformal prediction (CP). The Shapley Additive Explanations method was used to explain the model.

RESULTS

The multiclass gradient boosting machine was identified as the best-performing model with good discrimination and calibration performance in both validation cohorts. The mean area under the receiver operating characteristic curve with SD was 0.906 (0.018) for rapid death, 0.843 (0.008) for recovery, and 0.807 (0.010) for persistent ill in the internal validation cohort. In the external validation cohort, the mean area under the receiver operating characteristic curve (SD) was 0.878 (0.003) for rapid death, 0.764 (0.008) for recovery, and 0.696 (0.007) for persistent ill. The maximum norepinephrine equivalence, total urine output, Acute Physiology Score III, mean systolic blood pressure, and the coefficient of variation of oxygen saturation contributed the most. Compared to the model without CP, using the model with CP at a mixed confidence approach reduced overall prediction errors by 27.6% (n=62) and 30.7% (n=412) in the internal and external validation cohorts, respectively, as well as enabled the identification of more potentially persistent ill patients.

CONCLUSIONS

The implementation of our model has the potential to reduce heterogeneity and enroll more homogeneous patients in sepsis clinical trials. The use of CP for estimating the uncertainty of the model outputs allows for a more comprehensive understanding of the model's reliability and assists in making informed decisions based on the predicted outcomes.

Improving sepsis prediction in intensive care with SepsisAI: A clinical decision support system with a focus on minimizing false alarms

Prediction of sepsis using machine-learning approaches has recently gained traction. However, the lack of translation of these algorithms into clinical routine remains a major issue. Existing early sepsis detection methods are either based on the older definition of sepsis or do not accurately detect sepsis leading to the high frequency of false-positive alarms. This results in a well-known issue of clinicians' "alarm fatigue", leading to decreased responsiveness and identification, ultimately resulting in delayed clinical intervention. Hence, there is a fundamental, unmet need for a clinical decision system capable of accurate and timely sepsis diagnosis, running at the point of need. In this work, SepsisAI-a deep-learning algorithm based on long short-term memory (LSTM) networks was developed to predict the early onset of hospital-acquired sepsis in real-time for patients admitted to the ICU. The models are trained and validated with data from the PhysioNet Challenge, consisting of 40,336 patient data files from two healthcare systems: Beth Israel Deaconess Medical Center and Emory University Hospital. In the short term, the algorithm tracks frequently measured vital signs, sparsely available lab parameters, demographic features, and certain derived features for making predictions. A real-time alert system, which monitors the trajectory of the predictions, is developed on top of the deep-learning framework to minimize false alarms. On a balanced test dataset, the model achieves an AUROC, AUPRC, sensitivity, and specificity of 0.95, 0.96, 88.19%, and 96.75%, respectively at the patient level. In terms of lookahead time, the model issues a warning at a median of 6 hours (IQR 6 to 20 hours) and raises an alert at a median of 4 hours (IQR 2 to 5 hours) ahead of sepsis onset. Most importantly, the model achieves a false-alarm ratio of 3.18% for alerts, which is significantly less than other sepsis alarm systems. Additionally, on a disease prevalence-based test set, the algorithm reported similar outcomes with AUROC and AUPRC of 0.94 and 0.87, respectively, with sensitivity, and specificity of 97.05%, and 96.75%, respectively. The proposed algorithm might serve as a clinical decision support system to assist clinicians in the accurate and timely diagnosis of sepsis. With exceptionally high specificity and low false-alarm rate, this algorithm also helps mitigate the well-known issue of clinician alert fatigue arising from currently proposed sepsis alarm systems. Consequently, the algorithm partially addresses the challenges of successfully integrating machine-learning algorithms into routine clinical care.

Early Warning Systems for Critical Illness Outside the Intensive Care Unit

Early warning systems (EWSs) are designed and deployed to create a rapid assessment and response for patients with clinical deterioration outside the intensive care unit (ICU). These models incorporate patient-level data such as vital signs and laboratory values to detect or prevent adverse clinical events, such as vital signs and laboratories to allow detection and prevention of adverse clinical events such as cardiac arrest, intensive care transfer, or sepsis. The applicability, development, clinical utility, and general perception of EWS in clinical practice vary widely. Here, we review the field as it has grown from early vital sign-based scoring systems to contemporary multidimensional algorithms and predictive technologies for clinical decompensation outside the ICU.

Early identification of suspected serious infection among patients afebrile at initial presentation using neural network models and natural language processing: A development and external validation study in the emergency department

OBJECTIVE

To develop and externally validate models based on neural networks and natural language processing (NLP) to identify suspected serious infections in emergency department (ED) patients afebrile at initial presentation.

METHODS

This retrospective study included adults who visited the ED afebrile at initial presentation. We developed four models based on artificial neural networks to identify suspected serious infection. Patient demographics, vital signs, laboratory test results and information extracted from initial ED physician notes using term frequency-inverse document frequency were used as model variables. Models were trained and internally validated with data from one hospital and externally validated using data from a different hospital. Model discrimination was evaluated using area under the receiver operating characteristic curve (AUC) and 95% confidence intervals (CIs).

RESULTS

The training, internal validation, and external validation datasets comprised 150,699, 37,675, and 85,098 patients, respectively. The AUCs (95% CIs) for Models 1 (demographics + vital signs), 2 (demographics + vital signs + initial ED physician note), 3 (demographics + vital signs + laboratory tests), and 4 (demographics + vital signs + laboratory tests + initial ED physician note) in the internal validation dataset were 0.789 (0.782-0.796), 0.867 (0.862-0.872), 0.881 (0.876-0.887), and 0.911 (0.906-0.915), respectively. In the external validation dataset, the AUCs (95% CIs) of Models 1, 2, 3, and 4 were 0.824 (0.817-0.830), 0.895 (0.890-0.899), 0.879 (0.873-0.884), and 0.913 (0.909-0.917), respectively. Model 1 can be utilized immediately after ED triage, Model 2 can be utilized after the initial physician notes are recorded (median time from ED triage: 28 min), and Models 3 and 4 can be utilized after the initial laboratory tests are reported (median time from ED triage: 68 min).

CONCLUSIONS

We developed and validated models to identify suspected serious infection in the ED. Extracted information from initial ED physician notes using NLP contributed to increased model performance, permitting identification of suspected serious infection at early stages of ED visits.

A scoping review of machine learning for sepsis prediction- feature engineering strategies and model performance: a step towards explainability

BACKGROUND

Sepsis, an acute and potentially fatal systemic response to infection, significantly impacts global health by affecting millions annually. Prompt identification of sepsis is vital, as treatment delays lead to increased fatalities through progressive organ dysfunction. While recent studies have delved into leveraging Machine Learning (ML) for predicting sepsis, focusing on aspects such as prognosis, diagnosis, and clinical application, there remains a notable deficiency in the discourse regarding feature engineering. Specifically, the role of feature selection and extraction in enhancing model accuracy has been underexplored.

OBJECTIVES

This scoping review aims to fulfill two primary objectives: To identify pivotal features for predicting sepsis across a variety of ML models, providing valuable insights for future model development, and To assess model efficacy through performance metrics including AUROC, sensitivity, and specificity.

RESULTS

The analysis included 29 studies across diverse clinical settings such as Intensive Care Units (ICU), Emergency Departments, and others, encompassing 1,147,202 patients. The review highlighted the diversity in prediction strategies and timeframes. It was found that feature extraction techniques notably outperformed others in terms of sensitivity and AUROC values, thus indicating their critical role in improving sepsis prediction models.

CONCLUSION

Key dynamic indicators, including vital signs and critical laboratory values, are instrumental in the early detection of sepsis. Applying feature selection methods significantly boosts model precision, with models like Random Forest and XG Boost showing promising results. Furthermore, Deep Learning models (DL) reveal unique insights, spotlighting the pivotal role of feature engineering in sepsis prediction, which could greatly benefit clinical practice.

DEVELOPMENT AND VALIDATION OF A NOMOGRAM FOR PREDICTING 28-DAY IN-HOSPITAL MORTALITY IN SEPSIS PATIENTS BASED ON AN OPTIMIZED ACUTE PHYSIOLOGY AND CHRONIC HEALTH EVALUATION II SCORE

ABSTRACT

Purpose : The objective of this study is to establish a nomogram that correlates optimized Acute Physiology and Chronic Health Evaluation II (APACHE II) score with sepsis-related indicators, aiming to provide a robust model for early prediction of sepsis prognosis in clinical practice and serve as a valuable reference for improved diagnosis and treatment strategies. Methods : This retrospective study extracted sepsis patients meeting the inclusion criteria from the MIMIC-IV database to form the training group. An optimized APACHE II score integrated with relevant indicators was developed using a nomogram for predicting the prognosis of sepsis patients. External validation was conducted using data from the intensive care unit at Lanzhou University Second Hospital. Results : The study enrolled 1805 patients in the training cohort and 203 patients in the validation cohort. A multifactor analysis was conducted to identify factors affecting patient mortality within 28 days, resulting in the development of an optimized score by simplifying evaluation indicators from APACHE II score. The results showed that the optimized score (area under the ROC curve [AUC] = 0.715) had a higher area under receiver operating characteristic curve than Sequential Organ Failure Assessment score (AUC = 0.637) but slightly lower than APACHE II score (AUC = 0.720). Significant indicators identified through multifactor analysis included platelet count, total bilirubin level, albumin level, prothrombin time, activated partial thromboplastin time, mechanical ventilation use and renal replacement therapy use. These seven indicators were combined with optimized score to construct a nomogram based on these seven indicators. The nomogram demonstrated good clinical predictive value in both training cohort (AUC = 0.803) and validation cohort (AUC = 0.750). Calibration curves and decision curve analyses also confirmed its good predictive ability, surpassing the APACHE II score and Sequential Organ Failure Assessment score in identifying high-risk patients. Conclusions : The nomogram was established in this study using the MIMIC-IV database and validated with external data, demonstrating its robust discriminability, calibration, and clinical practicability for predicting 28-day mortality in sepsis patients. These findings aim to provide substantial support for clinicians' decision making.

Fall prediction in a quiet standing balance test via machine learning: Is it possible?

The elderly population is growing rapidly in the world and falls are becoming a big problem for society. Currently, clinical assessments of gait and posture include functional evaluations, objective, and subjective scales. They are considered the gold standard to indicate optimal mobility and performance individually, but their sensitivity and specificity are not good enough to predict who is at higher risk of falling. An innovative approach for fall prediction is the machine learning. Machine learning is a computer-science area that uses statistics and optimization methods in a large amount of data to make outcome predictions. Thus, to assess the performance of machine learning algorithms in classify participants by age, number of falls and falls frequency based on features extracted from a public database of stabilometric assessments. 163 participants (116 women and 47 men) between 18 and 85 years old, 44.0 to 75.9 kg mass, 140.0 to 189.8 cm tall, and 17.2 to 31.9 kg/m2 body mass index. Six different machine learning algorithms were tested for this classification, which included Logistic Regression, Linear Discriminant Analysis, K Nearest-neighbours, Decision Tree Classifier, Gaussian Naive Bayes and C-Support Vector Classification. The machine learning algorithms were applied in this database which has sociocultural, demographic, and health status information about participants. All algorithm models were able to classify the participants into young or old, but our main goal was not achieved, no model identified participants at high risk of falling. Our conclusion corroborates other works in the biomechanics field, arguing the static posturography, probably due to the low daily living activities specificity, does not have the desired effects in predicting the risk of falling. Further studies should focus on dynamic posturography to assess the risk of falls.

Application of AI in Sepsis: Citation Network Analysis and Evidence Synthesis

BACKGROUND

Artificial intelligence (AI) has garnered considerable attention in the context of sepsis research, particularly in personalized diagnosis and treatment. Conducting a bibliometric analysis of existing publications can offer a broad overview of the field and identify current research trends and future research directions.

OBJECTIVE

The objective of this study is to leverage bibliometric data to provide a comprehensive overview of the application of AI in sepsis.

METHODS

We conducted a search in the Web of Science Core Collection database to identify relevant articles published in English until August 31, 2023. A predefined search strategy was used, evaluating titles, abstracts, and full texts as needed. We used the Bibliometrix and VOSviewer tools to visualize networks showcasing the co-occurrence of authors, research institutions, countries, citations, and keywords.

RESULTS

A total of 259 relevant articles published between 2014 and 2023 (until August) were identified. Over the past decade, the annual publication count has consistently risen. Leading journals in this domain include Critical Care Medicine (17/259, 6.6%), Frontiers in Medicine (17/259, 6.6%), and Scientific Reports (11/259, 4.2%). The United States (103/259, 39.8%), China (83/259, 32%), United Kingdom (14/259, 5.4%), and Taiwan (12/259, 4.6%) emerged as the most prolific countries in terms of publications. Notable institutions in this field include the University of California System, Emory University, and Harvard University. The key researchers working in this area include Ritankar Das, Chris Barton, and Rishikesan Kamaleswaran. Although the initial period witnessed a relatively low number of articles focused on AI applications for sepsis, there has been a significant surge in research within this area in recent years (2014-2023).

CONCLUSIONS

This comprehensive analysis provides valuable insights into AI-related research conducted in the field of sepsis, aiding health care policy makers and researchers in understanding the potential of AI and formulating effective research plans. Such analysis serves as a valuable resource for determining the advantages, sustainability, scope, and potential impact of AI models in sepsis.

Predicting septic shock in patients with sepsis at emergency department triage using systolic and diastolic shock index

INTRODUCTION

Identifying patients with at a high risk of progressing to septic shock is essential. Due to systemic vasodilation in the pathophysiology of septic shock, the use of diastolic blood pressure (DBP) has emerged. We hypothesized that the initial shock index (SI) and diastolic SI (DSI) at the emergency department (ED) triage can predict septic shock.

METHOD

This observational study used the prospectively collected sepsis registry. The primary outcome was progression to septic shock. Secondary outcomes were the time to vasopressor requirement, vasopressor dose, and severity according to SI and DSI. Patients were classified by tertiles according to the first principal component of shock index and diastolic shock index.

RESULTS

A total of 1267 patients were included in the analysis. The area under the receiver operating characteristic curve (AUC) for predicting progression to septic shock for DSI was 0.717, while that for SI was 0.707. The AUC for predicting progression to septic shock for DSI and SI were significantly higher than those for conventional early warning scores. Middle tertile showed adjusted Odd ratio (aOR) of 1.448 (95% CI 1.074-1.953), and that of upper tertile showed 3.704 (95% CI 2.299-4.111).

CONCLUSION

The SI and DSI were significant predictors of progression to septic shock. Our findings suggest an association between DSI and vasopressor requirement. We propose stratifying lower tertile as being at low risk, middle tertile as being at intermediate risk, and upper tertile as being at high risk of progression to septic shock. This system can be applied simply at the ED triage.

Machine Learning Interpretability Methods to Characterize the Importance of Hematologic Biomarkers in Prognosticating Patients with Suspected Infection

Early detection of sepsis in patients admitted to the emergency department (ED) is an important clinical objective as early identification and treatment can help reduce morbidity and mortality rate of 20% or higher. Hematologic changes during sepsis-associated organ dysfunction are well established and a new biomarker called Monocyte Distribution Width (MDW) has been recently approved by the US Food and Drug Administration for sepsis. However, MDW, which quantifies monocyte activation in sepsis patients, is not a routinely reported parameter and it requires specialized proprietary laboratory equipment. Further, the relative importance of MDW as compared to other routinely available hematologic parameters and vital signs has not been studied, which makes it difficult for resource constrained hospital systems to make informed decisions in this regard. To address this issue, we analyzed data from a cohort of ED patients (n=10,229) admitted to a large regional safety-net hospital in Cleveland, Ohio with suspected infection who later developed poor outcomes associated with sepsis. We developed a new analytical framework consisting of seven data models and an ensemble of high accuracy machine learning (ML) algorithms (accuracy values ranging from 0.83 to 0.90) for the prediction of outcomes more common in sepsis than uncomplicated infection (3-day intensive care unit stay or death). To characterize the contributions of individual hematologic parameters, we applied the Local Interpretable Model-Agnostic Explanation (LIME) and Shapley Additive Value (SHAP) interpretability methods to the high accuracy ML algorithms. The ML interpretability results were consistent in their findings that the value of MDW is grossly attenuated in the presence of other routinely reported hematologic parameters and vital signs data. Further, this study for the first time shows that complete blood count with differential (CBC-DIFF) together with vital signs data can be used as a substitute for MDW in high accuracy ML algorithms to screen for poor outcomes associated with sepsis.

Surviving Sepsis Campaign Research Priorities 2023

OBJECTIVES

To identify research priorities in the management, epidemiology, outcome, and pathophysiology of sepsis and septic shock.

DESIGN

Shortly after publication of the most recent Surviving Sepsis Campaign Guidelines, the Surviving Sepsis Research Committee, a multiprofessional group of 16 international experts representing the European Society of Intensive Care Medicine and the Society of Critical Care Medicine, convened virtually and iteratively developed the article and recommendations, which represents an update from the 2018 Surviving Sepsis Campaign Research Priorities.

METHODS

Each task force member submitted five research questions on any sepsis-related subject. Committee members then independently ranked their top three priorities from the list generated. The highest rated clinical and basic science questions were developed into the current article.

RESULTS

A total of 81 questions were submitted. After merging similar questions, there were 34 clinical and ten basic science research questions submitted for voting. The five top clinical priorities were as follows: 1) what is the best strategy for screening and identification of patients with sepsis, and can predictive modeling assist in real-time recognition of sepsis? 2) what causes organ injury and dysfunction in sepsis, how should it be defined, and how can it be detected? 3) how should fluid resuscitation be individualized initially and beyond? 4) what is the best vasopressor approach for treating the different phases of septic shock? and 5) can a personalized/precision medicine approach identify optimal therapies to improve patient outcomes? The five top basic science priorities were as follows: 1) How can we improve animal models so that they more closely resemble sepsis in humans? 2) What outcome variables maximize correlations between human sepsis and animal models and are therefore most appropriate to use in both? 3) How does sepsis affect the brain, and how do sepsis-induced brain alterations contribute to organ dysfunction? How does sepsis affect interactions between neural, endocrine, and immune systems? 4) How does the microbiome affect sepsis pathobiology? 5) How do genetics and epigenetics influence the development of sepsis, the course of sepsis and the response to treatments for sepsis?

CONCLUSIONS

Knowledge advances in multiple clinical domains have been incorporated in progressive iterations of the Surviving Sepsis Campaign guidelines, allowing for evidence-based recommendations for short- and long-term management of sepsis. However, the strength of existing evidence is modest with significant knowledge gaps and mortality from sepsis remains high. The priorities identified represent a roadmap for research in sepsis and septic shock.

Impact of a deep learning sepsis prediction model on quality of care and survival

Sepsis remains a major cause of mortality and morbidity worldwide. Algorithms that assist with the early recognition of sepsis may improve outcomes, but relatively few studies have examined their impact on real-world patient outcomes. Our objective was to assess the impact of a deep-learning model (COMPOSER) for the early prediction of sepsis on patient outcomes. We completed a before-and-after quasi-experimental study at two distinct Emergency Departments (EDs) within the UC San Diego Health System. We included 6217 adult septic patients from 1/1/2021 through 4/30/2023. The exposure tested was a nurse-facing Best Practice Advisory (BPA) triggered by COMPOSER. In-hospital mortality, sepsis bundle compliance, 72-h change in sequential organ failure assessment (SOFA) score following sepsis onset, ICU-free days, and the number of ICU encounters were evaluated in the pre-intervention period (705 days) and the post-intervention period (145 days). The causal impact analysis was performed using a Bayesian structural time-series approach with confounder adjustments to assess the significance of the exposure at the 95% confidence level. The deployment of COMPOSER was significantly associated with a 1.9% absolute reduction (17% relative decrease) in in-hospital sepsis mortality (95% CI, 0.3%-3.5%), a 5.0% absolute increase (10% relative increase) in sepsis bundle compliance (95% CI, 2.4%-8.0%), and a 4% (95% CI, 1.1%-7.1%) reduction in 72-h SOFA change after sepsis onset in causal inference analysis. This study suggests that the deployment of COMPOSER for early prediction of sepsis was associated with a significant reduction in mortality and a significant increase in sepsis bundle compliance.

The critical role of neutrophil-endothelial cell interactions in sepsis: new synergistic approaches employing organ-on-chip, omics, immune cell phenotyping and in silico modeling to identify new therapeutics

Sepsis is a global health concern accounting for more than 1 in 5 deaths worldwide. Sepsis is now defined as life-threatening organ dysfunction caused by a dysregulated host response to infection. Sepsis can develop from bacterial (gram negative or gram positive), fungal or viral (such as COVID) infections. However, therapeutics developed in animal models and traditional in vitro sepsis models have had little success in clinical trials, as these models have failed to fully replicate the underlying pathophysiology and heterogeneity of the disease. The current understanding is that the host response to sepsis is highly diverse among patients, and this heterogeneity impacts immune function and response to infection. Phenotyping immune function and classifying sepsis patients into specific endotypes is needed to develop a personalized treatment approach. Neutrophil-endothelium interactions play a critical role in sepsis progression, and increased neutrophil influx and endothelial barrier disruption have important roles in the early course of organ damage. Understanding the mechanism of neutrophil-endothelium interactions and how immune function impacts this interaction can help us better manage the disease and lead to the discovery of new diagnostic and prognosis tools for effective treatments. In this review, we will discuss the latest research exploring how in silico modeling of a synergistic combination of new organ-on-chip models incorporating human cells/tissue, omics analysis and clinical data from sepsis patients will allow us to identify relevant signaling pathways and characterize specific immune phenotypes in patients. Emerging technologies such as machine learning can then be leveraged to identify druggable therapeutic targets and relate them to immune phenotypes and underlying infectious agents. This synergistic approach can lead to the development of new therapeutics and the identification of FDA approved drugs that can be repurposed for the treatment of sepsis.

Information displays for automated surveillance algorithms of in-hospital patient deterioration: a scoping review

OBJECTIVE

Surveillance algorithms that predict patient decompensation are increasingly integrated with clinical workflows to help identify patients at risk of in-hospital deterioration. This scoping review aimed to identify the design features of the information displays, the types of algorithm that drive the display, and the effect of these displays on process and patient outcomes.

MATERIALS AND METHODS

The scoping review followed Arksey and O'Malley's framework. Five databases were searched with dates between January 1, 2009 and January 26, 2022. Inclusion criteria were: participants-clinicians in inpatient settings; concepts-intervention as deterioration information displays that leveraged automated AI algorithms; comparison as usual care or alternative displays; outcomes as clinical, workflow process, and usability outcomes; and context as simulated or real-world in-hospital settings in any country. Screening, full-text review, and data extraction were reviewed independently by 2 researchers in each step. Display categories were identified inductively through consensus.

RESULTS

Of 14 575 articles, 64 were included in the review, describing 61 unique displays. Forty-one displays were designed for specific deteriorations (eg, sepsis), 24 provided simple alerts (ie, text-based prompts without relevant patient data), 48 leveraged well-accepted score-based algorithms, and 47 included nurses as the target users. Only 1 out of the 10 randomized controlled trials reported a significant effect on the primary outcome.

CONCLUSIONS

Despite significant advancements in surveillance algorithms, most information displays continue to leverage well-understood, well-accepted score-based algorithms. Users' trust, algorithmic transparency, and workflow integration are significant hurdles to adopting new algorithms into effective decision support tools.

Explainable hierarchical clustering for patient subtyping and risk prediction

We present a pipeline in which machine learning techniques are used to automatically identify and evaluate subtypes of hospital patients admitted between 2017 and 2021 in a large UK teaching hospital. Patient clusters are determined using routinely collected hospital data, such as those used in the UK's National Early Warning Score 2 (NEWS2). An iterative, hierarchical clustering process was used to identify the minimum set of relevant features for cluster separation. With the use of state-of-the-art explainability techniques, the identified subtypes are interpreted and assigned clinical meaning, illustrating their robustness. In parallel, clinicians assessed intracluster similarities and intercluster differences of the identified patient subtypes within the context of their clinical knowledge. For each cluster, outcome prediction models were trained and their forecasting ability was illustrated against the NEWS2 of the unclustered patient cohort. These preliminary results suggest that subtype models can outperform the established NEWS2 method, providing improved prediction of patient deterioration. By considering both the computational outputs and clinician-based explanations in patient subtyping, we aim to highlight the mutual benefit of combining machine learning techniques with clinical expertise.

From immunology to artificial intelligence: revolutionizing latent tuberculosis infection diagnosis with machine learning

Latent tuberculosis infection (LTBI) has become a major source of active tuberculosis (ATB). Although the tuberculin skin test and interferon-gamma release assay can be used to diagnose LTBI, these methods can only differentiate infected individuals from healthy ones but cannot discriminate between LTBI and ATB. Thus, the diagnosis of LTBI faces many challenges, such as the lack of effective biomarkers from Mycobacterium tuberculosis (MTB) for distinguishing LTBI, the low diagnostic efficacy of biomarkers derived from the human host, and the absence of a gold standard to differentiate between LTBI and ATB. Sputum culture, as the gold standard for diagnosing tuberculosis, is time-consuming and cannot distinguish between ATB and LTBI. In this article, we review the pathogenesis of MTB and the immune mechanisms of the host in LTBI, including the innate and adaptive immune responses, multiple immune evasion mechanisms of MTB, and epigenetic regulation. Based on this knowledge, we summarize the current status and challenges in diagnosing LTBI and present the application of machine learning (ML) in LTBI diagnosis, as well as the advantages and limitations of ML in this context. Finally, we discuss the future development directions of ML applied to LTBI diagnosis.

Effects of machine learning-based clinical decision support systems on decision-making, care delivery, and patient outcomes: a scoping review

OBJECTIVE

This study aims to summarize the research literature evaluating machine learning (ML)-based clinical decision support (CDS) systems in healthcare settings.

MATERIALS AND METHODS

We conducted a review in accordance with the PRISMA-ScR (Preferred Reporting Items for Systematic Reviews and Meta Analyses extension for Scoping Review). Four databases, including PubMed, Medline, Embase, and Scopus were searched for studies published from January 2016 to April 2021 evaluating the use of ML-based CDS in clinical settings. We extracted the study design, care setting, clinical task, CDS task, and ML method. The level of CDS autonomy was examined using a previously published 3-level classification based on the division of clinical tasks between the clinician and CDS; effects on decision-making, care delivery, and patient outcomes were summarized.

RESULTS

Thirty-two studies evaluating the use of ML-based CDS in clinical settings were identified. All were undertaken in developed countries and largely in secondary and tertiary care settings. The most common clinical tasks supported by ML-based CDS were image recognition and interpretation (n = 12) and risk assessment (n = 9). The majority of studies examined assistive CDS (n = 23) which required clinicians to confirm or approve CDS recommendations for risk assessment in sepsis and for interpreting cancerous lesions in colonoscopy. Effects on decision-making, care delivery, and patient outcomes were mixed.

CONCLUSION

ML-based CDS are being evaluated in many clinical areas. There remain many opportunities to apply and evaluate effects of ML-based CDS on decision-making, care delivery, and patient outcomes, particularly in resource-constrained settings.

A Systematic Review of the Barriers to the Implementation of Artificial Intelligence in Healthcare

Artificial intelligence (AI) is expected to improve healthcare outcomes by facilitating early diagnosis, reducing the medical administrative burden, aiding drug development, personalising medical and oncological management, monitoring healthcare parameters on an individual basis, and allowing clinicians to spend more time with their patients. In the post-pandemic world where there is a drive for efficient delivery of healthcare and manage long waiting times for patients to access care, AI has an important role in supporting clinicians and healthcare systems to streamline the care pathways and provide timely and high-quality care for the patients. Despite AI technologies being used in healthcare for some decades, and all the theoretical potential of AI, the uptake in healthcare has been uneven and slower than anticipated and there remain a number of barriers, both overt and covert, which have limited its incorporation. This literature review highlighted barriers in six key areas: ethical, technological, liability and regulatory, workforce, social, and patient safety barriers. Defining and understanding the barriers preventing the acceptance and implementation of AI in the setting of healthcare will enable clinical staff and healthcare leaders to overcome the identified hurdles and incorporate AI technologies for the benefit of patients and clinical staff.

The Learning Electronic Health Record

Electronic medical records (EMRs) constitute the electronic version of all medical information included in a patient's paper chart. The electronic health record (EHR) technology has witnessed massive expansion in developed countries and to a lesser extent in underresourced countries during the last 2 decades. We will review factors leading to this expansion, how the emergence of EHRs is affecting several health-care stakeholders; some of the growing pains associated with EHRs with a particular emphasis on the delivery of care to the critically ill; and ongoing developments on the path to improve the quality of research, health-care delivery, and stakeholder satisfaction.

Clinician Trust in Artificial Intelligence: What is Known and How Trust Can Be Facilitated

Predictive analytics based on artificial intelligence (AI) offer clinicians the opportunity to leverage big data available in electronic health records (EHR) to improve clinical decision-making, and thus patient outcomes. Despite this, many barriers exist to facilitating trust between clinicians and AI-based tools, limiting its current impact. Potential solutions are available at both the local and national level. It will take a broad and diverse coalition of stakeholders, from health-care systems, EHR vendors, and clinical educators to regulators, researchers and the patient community, to help facilitate this trust so that the promise of AI in health care can be realized.

Predictive Modeling Using Artificial Intelligence and Machine Learning Algorithms on Electronic Health Record Data: Advantages and Challenges

The rapid adoption of electronic health record (EHR) systems in US hospitals from 2008 to 2014 produced novel data elements for analysis. Concurrent innovations in computing architecture and machine learning (ML) algorithms have made rapid consumption of health data feasible and a powerful engine for clinical innovation. In critical care research, the net convergence of these trends has resulted in an exponential increase in outcome prediction research. In the following article, we explore the history of outcome prediction in the intensive care unit (ICU), the growing use of EHR data, and the rise of artificial intelligence and ML (AI) in critical care.

Predicting sepsis onset in ICU using machine learning models: a systematic review and meta-analysis

BACKGROUND

Sepsis is a life-threatening condition caused by an abnormal response of the body to infection and imposes a significant health and economic burden worldwide due to its high mortality rate. Early recognition of sepsis is crucial for effective treatment. This study aimed to systematically evaluate the performance of various machine learning models in predicting the onset of sepsis.

METHODS

We conducted a comprehensive search of the Cochrane Library, PubMed, Embase, and Web of Science databases, covering studies from database inception to November 14, 2022. We used the PROBAST tool to assess the risk of bias. We calculated the predictive performance for sepsis onset using the C-index and accuracy. We followed the PRISMA guidelines for this study.

RESULTS

We included 23 eligible studies with a total of 4,314,145 patients and 26 different machine learning models. The most frequently used models in the studies were random forest (n = 9), extreme gradient boost (n = 7), and logistic regression (n = 6) models. The random forest (test set n = 9, acc = 0.911) and extreme gradient boost (test set n = 7, acc = 0.957) models were the most accurate based on our analysis of the predictive performance. In terms of the C-index outcome, the random forest (n = 6, acc = 0.79) and extreme gradient boost (n = 7, acc = 0.83) models showed the highest performance.

CONCLUSION

Machine learning has proven to be an effective tool for predicting sepsis at an early stage. However, to obtain more accurate results, additional machine learning methods are needed. In our research, we discovered that the XGBoost and random forest models exhibited the best predictive performance and were most frequently utilized for predicting the onset of sepsis.

TRIAL REGISTRATION

CRD42022384015.

Elaborating the potential of Artificial Intelligence in automated CAR-T cell manufacturing

This paper discusses the challenges of producing CAR-T cells for cancer treatment and the potential for Artificial Intelligence (AI) for its improvement. CAR-T cell therapy was approved in 2018 as the first Advanced Therapy Medicinal Product (ATMP) for treating acute leukemia and lymphoma. ATMPs are cell- and gene-based therapies that show great promise for treating various cancers and hereditary diseases. While some new ATMPs have been approved, ongoing clinical trials are expected to lead to the approval of many more. However, the production of CAR-T cells presents a significant challenge due to the high costs associated with the manufacturing process, making the therapy very expensive (approx. $400,000). Furthermore, autologous CAR-T therapy is limited to a make-to-order approach, which makes scaling economical production difficult. First attempts are being made to automate this multi-step manufacturing process, which will not only directly reduce the high manufacturing costs but will also enable comprehensive data collection. AI technologies have the ability to analyze this data and convert it into knowledge and insights. In order to exploit these opportunities, this paper analyses the data potential in the automated CAR-T production process and creates a mapping to the capabilities of AI applications. The paper explores the possible use of AI in analyzing the data generated during the automated process and its capabilities to further improve the efficiency and cost-effectiveness of CAR-T cell production.

An Individualized Nomogram for Predicting Mortality Risk of Septic Shock Patients During Hospitalization: A ten Years Retrospective Analysis

Purpose

We intend to develop a nomogram for predicting the mortality risk of hospitalized septic shock patients.

Patients and Methods

Data were collected from patients hospitalized with septic shock in Affiliated Dongyang Hospital of Wenzhou Medical University in China, over 10 years between January 2013 and January 2023. The eligible study participants were divided into modeling and validation groups. Factors independently related to the mortality in the modeling group were obtained by stepwise regression analysis. A logistic regression model and a nomogram were built. The model was evaluated based on the discrimination power (the area under the curve of the receiver operating characteristic, AUC), the calibration degree and decision curve analysis. In the validation group, the discrimination powers of the logistic regression model, the sequential organ failure assessment (SOFA) scoring model and machine learning model were compared.

Results

A total of 1253 patients, including 878 patients in the modeling group and 375 patients in the validation group, were included in this study. Age, respiratory failure, serum cholinesterase, lactic acid, blood phosphorus, blood magnesium, total bilirubin, and pH were independent risk factors related to the mortality risk of septic shock. The AUCs of the prediction model for the modeling and validation groups were 0.881 and 0.868, respectively. The models had a good calibration degree and clinical applicability. The AUC of the SOFA model for the validation population was 0.799, significantly lower than that of our model. The AUCs of the random forest and ensemble models were 0.865 and 0.863, respectively, comparable to that of our logistical prediction model.

Conclusion

The model established in this study can effectively predict the mortality risk in patients hospitalized with septic shock. Thus, the model could be used clinically to determine the best therapy or management for patients with septic shock.

Clinical Decision Support System to Detect the Occurrence of Ventilator-Associated Pneumonia in Pediatric Intensive Care

OBJECTIVES

Ventilator-associated pneumonia (VAP) is a severe care-related disease. The Centers for Disease Control defined the diagnosis criteria; however, the pediatric criteria are mainly subjective and retrospective. Clinical decision support systems have recently been developed in healthcare to help the physician to be more accurate for the early detection of severe pathology. We aimed at developing a predictive model to provide early diagnosis of VAP at the bedside in a pediatric intensive care unit (PICU).

METHODS

We performed a retrospective single-center study at a tertiary-care pediatric teaching hospital. All patients treated by invasive mechanical ventilation between September 2013 and October 2019 were included. Data were collected in the PICU electronic medical record and high-resolution research database. Development of the clinical decision support was then performed using open-access R software (Version 3.6.1®).

MEASUREMENTS AND MAIN RESULTS

In total, 2077 children were mechanically ventilated. We identified 827 episodes with almost 48 h of mechanical invasive ventilation and 77 patients who suffered from at least one VAP event. We split our database at the patient level in a training set of 461 patients free of VAP and 45 patients with VAP and in a testing set of 199 patients free of VAP and 20 patients with VAP. The Imbalanced Random Forest model was considered as the best fit with an area under the ROC curve from fitting the Imbalanced Random Forest model on the testing set being 0.82 (95% CI: (0.71, 0.93)). An optimal threshold of 0.41 gave a sensitivity of 79.7% and a specificity of 72.7%, with a positive predictive value (PPV) of 9% and a negative predictive value of 99%, and with an accuracy of 79.5% (95% CI: (0.77, 0.82)).

CONCLUSIONS

Using machine learning, we developed a clinical predictive algorithm based on clinical data stored prospectively in a database. The next step will be to implement the algorithm in PICUs to provide early, automatic detection of ventilator-associated pneumonia.

Using machine learning for process improvement in sepsis management

INTRODUCTION

In the U.S., sepsis afflicts 1.7 million adults, causing 270,000 deaths each year. Early detection of sepsis could decrease the number of deaths by 92,000 annually and decrease hospital expenditures by 1.5 billion USD. Few prior studies and reviews have presented a holistic understanding of the relationship between machine learning and existing process improvement measures. This study, in addition to discussing machine learning and existing process improvements measures, elaborates on the disadvantages and the barriers to integrating machine learning into the clinic. This article synthesizes previous studies to educate healthcare professionals on effectively managing sepsis by leveraging the benefits of machine learning.

METHODS

This study used the PubMed database. Search terms include sepsis antibiotics, sepsis process improvement, sepsis machine learning. Our search criteria included previous studies published between January 1, 2017, and February 1, 2022.

RESULTS/DISCUSSION

Although machine learning algorithms have better predictive capabilities, their effectiveness in the clinical setting is limited as studies show mixed results because the medical staff often fails to intervene. To overcome poor interventional response, clinicians need to work with the facility's IT department to ensure integration into clinical workflow and minimize alert-fatigue. Algorithms should enhance the productivity of clinical teams, not attempt to replace them entirely.

CONCLUSION

Hospitals can employ process improvement measures that effectively utilize machine learning algorithms to ensure integration into clinical workflows. Healthcare professionals can utilize workflow tools in addition to the predictive capabilities of machine learning to enhance clinical decisions in sepsis.

EARLY PREDICTION OF UNEXPECTED LATENT SHOCK IN THE EMERGENCY DEPARTMENT USING VITAL SIGNS

ABSTRACT

Objective/Introduction : Sequential vital-sign information and trends in vital signs are useful for predicting changes in patient state. This study aims to predict latent shock by observing sequential changes in patient vital signs. Methods : The dataset for this retrospective study contained a total of 93,194 emergency department (ED) visits from January 1, 2016, and December 31, 2020, and Medical Information Mart for Intensive Care (MIMIC)-IV-ED data. We further divided the data into training and validation datasets by random sampling without replacement at a 7:3 ratio. We carried out external validation with MIMIC-IV-ED. Our prediction model included logistic regression (LR), random forest (RF) classifier, a multilayer perceptron (MLP), and a recurrent neural network (RNN). To analyze the model performance, we used area under the receiver operating characteristic curve (AUROC). Results : Data of 89,250 visits of patients who met prespecified criteria were used to develop a latent-shock prediction model. Data of 142,250 patient visits from MIMIC-IV-ED satisfying the same inclusion criteria were used for external validation of the prediction model. The AUROC values of prediction for latent shock were 0.822, 0.841, 0.852, and 0.830 with RNN, MLP, RF, and LR methods, respectively, at 3 h before latent shock. This is higher than the shock index or adjusted shock index. Conclusion : We developed a latent shock prediction model based on 24 h of vital-sign sequence that changed with time and predicted the results by individual.

Clinical Informatics in Critical Care Medicine

Continuous monitoring and treatment of patients in intensive care units generates vast amounts of data. Critical Care Medicine clinicians incorporate this continuously evolving data to make split-second, life or death decisions for management of these patients. Despite the abundance of data, it can be challenging to consider every accessible data point when making the quick decisions necessary at the point of care. Consequently, Clinical Informatics offers a natural partnership to improve the care for critically ill patients. The last two decades have seen a significant evolution in the role of Clinical Informatics in Critical Care Medicine. In this review, we will discuss how Clinical Informatics improves the care of critically ill patients by enhancing not only data collection and visualization but also bedside medical decision making. We will further discuss the evolving role of machine learning algorithms in Clinical Informatics as it pertains to Critical Care Medicine.

Machine Learning-Based Early Prediction of Sepsis Using Electronic Health Records: A Systematic Review

BACKGROUND

Sepsis, a life-threatening infection-induced inflammatory condition, has significant global health impacts. Timely detection is crucial for improving patient outcomes as sepsis can rapidly progress to severe forms. The application of machine learning (ML) and deep learning (DL) to predict sepsis using electronic health records (EHRs) has gained considerable attention for timely intervention.

METHODS

PubMed, IEEE Xplore, Google Scholar, and Scopus were searched for relevant studies. All studies that used ML/DL to detect or early-predict the onset of sepsis in the adult population using EHRs were considered. Data were extracted and analyzed from all studies that met the criteria and were also evaluated for their quality.

RESULTS

This systematic review examined 1942 articles, selecting 42 studies while adhering to strict criteria. The chosen studies were predominantly retrospective (n = 38) and spanned diverse geographic settings, with a focus on the United States. Different datasets, sepsis definitions, and prevalence rates were employed, necessitating data augmentation. Heterogeneous parameter utilization, diverse model distribution, and varying quality assessments were observed. Longitudinal data enabled early sepsis prediction, and quality criteria fulfillment varied, with inconsistent funding-article quality correlation.

CONCLUSIONS

This systematic review underscores the significance of ML/DL methods for sepsis detection and early prediction through EHR data.

Predicting sepsis onset using a machine learned causal probabilistic network algorithm based on electronic health records data

Sepsis is a leading cause of mortality and early identification improves survival. With increasing digitalization of health care data automated sepsis prediction models hold promise to aid in prompt recognition. Most previous studies have focused on the intensive care unit (ICU) setting. Yet only a small proportion of sepsis develops in the ICU and there is an apparent clinical benefit to identify patients earlier in the disease trajectory. In this cohort of 82,852 hospital admissions and 8038 sepsis episodes classified according to the Sepsis-3 criteria, we demonstrate that a machine learned score can predict sepsis onset within 48 h using sparse routine electronic health record data outside the ICU. Our score was based on a causal probabilistic network model-SepsisFinder-which has similarities with clinical reasoning. A prediction was generated hourly on all admissions, providing a new variable was registered. Compared to the National Early Warning Score (NEWS2), which is an established method to identify sepsis, the SepsisFinder triggered earlier and had a higher area under receiver operating characteristic curve (AUROC) (0.950 vs. 0.872), as well as area under precision-recall curve (APR) (0.189 vs. 0.149). A machine learning comparator based on a gradient-boosting decision tree model had similar AUROC (0.949) and higher APR (0.239) than SepsisFinder but triggered later than both NEWS2 and SepsisFinder. The precision of SepsisFinder increased if screening was restricted to the earlier admission period and in episodes with bloodstream infection. Furthermore, the SepsisFinder signaled median 5.5 h prior to antibiotic administration. Identifying a high-risk population with this method could be used to tailor clinical interventions and improve patient care.

Two-stage learning-based prediction of bronchopulmonary dysplasia in very low birth weight infants: a nationwide cohort study

Introduction

The aim of this study is to develop an enhanced machine learning-based prediction models for bronchopulmonary dysplasia (BPD) and its severity through a two-stage approach integrated with the duration of respiratory support (RSd) using prenatal and early postnatal variables from a nationwide very low birth weight (VLBW) infant cohort.

Methods

We included 16,384 VLBW infants admitted to the neonatal intensive care unit (NICU) of the Korean Neonatal Network (KNN), a nationwide VLBW infant registry (2013-2020). Overall, 45 prenatal and early perinatal clinical variables were selected. A multilayer perceptron (MLP)-based network analysis, which was recently introduced to predict diseases in preterm infants, was used for modeling and a stepwise approach. Additionally, we applied a complementary MLP network and established new BPD prediction models (PMbpd). The performances of the models were compared using the area under the receiver operating characteristic curve (AUROC) values. The Shapley method was used to determine the contribution of each variable.

Results

We included 11,177 VLBW infants (3,724 without BPD (BPD 0), 3,383 with mild BPD (BPD 1), 1,375 with moderate BPD (BPD 2), and 2,695 with severe BPD (BPD 3) cases). Compared to conventional machine learning (ML) models, our PMbpd and two-stage PMbpd with RSd (TS-PMbpd) model outperformed both binary (0 vs. 1,2,3; 0,1 vs. 2,3; 0,1,2 vs. 3) and each severity (0 vs. 1 vs. 2 vs. 3) prediction (AUROC = 0.895 and 0.897, 0.824 and 0.825, 0.828 and 0.823, 0.783, and 0.786, respectively). GA, birth weight, and patent ductus arteriosus (PDA) treatment were significant variables for the occurrence of BPD. Birth weight, low blood pressure, and intraventricular hemorrhage were significant for BPD ≥2, birth weight, low blood pressure, and PDA ligation for BPD ≥3. GA, birth weight, and pulmonary hypertension were the principal variables that predicted BPD severity in VLBW infants.

Conclusions

We developed a new two-stage ML model reflecting crucial BPD indicators (RSd) and found significant clinical variables for the early prediction of BPD and its severity with high predictive accuracy. Our model can be used as an adjunctive predictive model in the practical NICU field.

Risk factors and machine learning prediction models for bronchopulmonary dysplasia severity in the Chinese population

BACKGROUND

Bronchopulmonary dysplasia (BPD) is a common chronic lung disease in extremely preterm neonates. The outcome and clinical burden vary dramatically according to severity. Although some prediction tools for BPD exist, they seldom pay attention to disease severity and are based on populations in developed countries. This study aimed to develop machine learning prediction models for BPD severity based on selected clinical factors in a Chinese population.

METHODS

In this retrospective, single-center study, we included patients with a gestational age < 32 weeks who were diagnosed with BPD in our neonatal intensive care unit from 2016 to 2020. We collected their clinical information during the maternal, birth and early postnatal periods. Risk factors were selected through univariable and ordinal logistic regression analyses. Prediction models based on logistic regression (LR), gradient boosting decision tree, XGBoost (XGB) and random forest (RF) models were implemented and assessed by the area under the receiver operating characteristic curve (AUC).

RESULTS

We ultimately included 471 patients (279 mild, 147 moderate, and 45 severe cases). On ordinal logistic regression, gestational diabetes mellitus, initial fraction of inspiration O2 value, invasive ventilation, acidosis, hypochloremia, C-reactive protein level, patent ductus arteriosus and Gram-negative respiratory culture were independent risk factors for BPD severity. All the XGB, LR and RF models (AUC = 0.85, 0.86 and 0.84, respectively) all had good performance.

CONCLUSIONS

We found risk factors for BPD severity in our population and developed machine learning models based on them. The models have good performance and can be used to aid in predicting BPD severity in the Chinese population.

Application Prospect of the SOFA Score and Related Modification Research Progress in Sepsis

In 2016, the SOFA score was proposed as the main evaluation system for diagnosis in the definition of sepsis 3.0, and the SOFA score has become a new research focus in sepsis. Some people are skeptical about diagnosing sepsis using the SOFA score. Experts and scholars from different regions have proposed different, modified versions of SOFA score to make up for the related problems with the use of the SOFA score in the diagnosis of sepsis. While synthesizing the different improved versions of SOFA proposed by experts and scholars in various regions, this paper also summarizes the relevant definitions of sepsis put forward in recent years to build a clear, improved application framework of SOFA score. In addition, the comparison between machine learning and SOFA scores related to sepsis is described and discussed in the article. Taken together, by summarizing the application of the improved SOFA score proposed in recent years in the related definition of sepsis, we believe that the SOFA score is still an effective means of diagnosing sepsis, but in the process of the continuous refinement and development of sepsis in the future, the SOFA score needs to be further refined and improved to provide more accurate coping strategies for different patient populations or application directions regarding sepsis. Against the big data background, machine learning has immeasurable value and significance, but its future applications should add more humanistic references and assistance.

Deployment of machine learning algorithms to predict sepsis: systematic review and application of the SALIENT clinical AI implementation framework

OBJECTIVE

To retrieve and appraise studies of deployed artificial intelligence (AI)-based sepsis prediction algorithms using systematic methods, identify implementation barriers, enablers, and key decisions and then map these to a novel end-to-end clinical AI implementation framework.

MATERIALS AND METHODS

Systematically review studies of clinically applied AI-based sepsis prediction algorithms in regard to methodological quality, deployment and evaluation methods, and outcomes. Identify contextual factors that influence implementation and map these factors to the SALIENT implementation framework.

RESULTS

The review identified 30 articles of algorithms applied in adult hospital settings, with 5 studies reporting significantly decreased mortality post-implementation. Eight groups of algorithms were identified, each sharing a common algorithm. We identified 14 barriers, 26 enablers, and 22 decision points which were able to be mapped to the 5 stages of the SALIENT implementation framework.

DISCUSSION

Empirical studies of deployed sepsis prediction algorithms demonstrate their potential for improving care and reducing mortality but reveal persisting gaps in existing implementation guidance. In the examined publications, key decision points reflecting real-word implementation experience could be mapped to the SALIENT framework and, as these decision points appear to be AI-task agnostic, this framework may also be applicable to non-sepsis algorithms. The mapping clarified where and when barriers, enablers, and key decisions arise within the end-to-end AI implementation process.

CONCLUSIONS

A systematic review of real-world implementation studies of sepsis prediction algorithms was used to validate an end-to-end staged implementation framework that has the ability to account for key factors that warrant attention in ensuring successful deployment, and which extends on previous AI implementation frameworks.

Evaluation of machine learning-based models for prediction of clinical deterioration: A systematic literature review

BACKGROUND AND OBJECTIVE

Early identification of patients at risk of deterioration can prevent life-threatening adverse events and shorten length of stay. Although there are numerous models applied to predict patient clinical deterioration, most are based on vital signs and have methodological shortcomings that are not able to provide accurate estimates of deterioration risk. The aim of this systematic review is to examine the effectiveness, challenges, and limitations of using machine learning (ML) techniques to predict patient clinical deterioration in hospital settings.

METHODS

A systematic review was performed in accordance with the Preferred Reporting Items for Systematic Reviews and meta-Analyses (PRISMA) guidelines using EMBASE, MEDLINE Complete, CINAHL Complete, and IEEExplore databases. Citation searching was carried out for studies that met inclusion criteria. Two reviewers used the inclusion/exclusion criteria to independently screen studies and extract data. To address any discrepancies in the screening process, the two reviewers discussed their findings and a third reviewer was consulted as needed to reach a consensus. Studies focusing on use of ML in predicting patient clinical deterioration that were published from inception to July 2022 were included.

RESULTS

A total of 29 primary studies that evaluated ML models to predict patient clinical deterioration were identified. After reviewing these studies, we found that 15 types of ML techniques have been employed to predict patient clinical deterioration. While six studies used a single technique exclusively, several others utilised a combination of classical techniques, unsupervised and supervised learning, as well as other novel techniques. Depending on which ML model was applied and the type of input features, ML models predicted outcomes with an area under the curve from 0.55 to 0.99.

CONCLUSIONS

Numerous ML methods have been employed to automate the identification of patient deterioration. Despite these advancements, there is still a need for further investigation to examine the application and effectiveness of these methods in real-world situations.

Early recognition of risk of critical adverse events based on deep neural decision gradient boosting

Introduction

Perioperative critical events will affect the quality of medical services and threaten the safety of patients. Using scientific methods to evaluate the perioperative risk of critical illness is of great significance for improving the quality of medical services and ensuring the safety of patients.

Method

At present, the traditional scoring system is mainly used to predict the score of critical illness, which is mainly dependent on the judgment of doctors. The result is affected by doctors' knowledge and experience, and the accuracy is difficult to guarantee and has a serious lag. Besides, the statistical prediction method based on pure data type do not make use of the patient's diagnostic text information and cannot identify comprehensive risk factor. Therefore, this paper combines the text features extracted by deep neural network with the pure numerical type features extracted by XGBOOST to propose a deep neural decision gradient boosting model. Supervised learning was used to train the risk prediction model to analyze the occurrence of critical illness during the perioperative period for early warning.

Results

We evaluated the proposed methods based on the real data of critical illness patients in one hospital from 2014 to 2018. The results showed that the critical disease risk prediction model based on multiple modes had faster convergence rate and better performance than the risk prediction model based on text data and pure data type.

Discussion

Based on the machine learning method and multi-modal data of patients, this paper built a prediction model for critical adverse events in patients, so that the risk of critical events can be predicted for any patient directly based on the preoperative and intraoperative characteristic data. At present, this work only classifies and predicts the occurrence of critical illness during or after operation based on the preoperative examination data of patients, but does not discuss the specific time when the patient was critical illness, which is also the direction of our future work.

Alert to Action: Implementing Artificial Intelligence-Driven Clinical Decision Support Tools for Sepsis

Background: Sepsis is the leading cause of mortality among hospitalized patients in our health care system and has been the target of major international initiatives such as the Surviving Sepsis Campaign championed by the Society of Critical Care Medicine and Get Ahead of Sepsis led by the Centers for Disease Control and Prevention. Methods: Our institution has strived to improve outcomes for patients by implementing a novel suite of integrated clinical decision support tools driven by a predictive learning algorithm in the electronic health record. The tools focus on sepsis multidisciplinary care using industry-standard heuristics of interface design to enhance usability and interaction. Results: Our novel clinical decision support tools demonstrated a higher level of interaction with a higher alert-to-action ratio compared to the average of all best practice alerts used at Ochsner Health (16.46% vs 8.4% to 12.1%). Conclusion: By using intuitive design strategies that encouraged users to complete best practice alerts and team-wide visualization of clinical decisions via a checklist, our clinical decision support tools for the detection and management of sepsis represent an improvement over legacy tools, and the results of this pilot may have implications beyond sepsis alerting.

A system theory based digital model for predicting the cumulative fluid balance course in intensive care patients

Background: Surgical interventions can cause severe fluid imbalances in patients undergoing cardiac surgery, affecting length of hospital stay and survival. Therefore, appropriate management of daily fluid goals is a key element of postoperative intensive care in these patients. Because fluid balance is influenced by a complex interplay of patient-, surgery- and intensive care unit (ICU)-specific factors, fluid prediction is difficult and often inaccurate. Methods: A novel system theory based digital model for cumulative fluid balance (CFB) prediction is presented using recorded patient fluid data as the sole parameter source by applying the concept of a transfer function. Using a retrospective dataset of n = 618 cardiac intensive care patients, patient-individual models were created and evaluated. RMSE analyses and error calculations were performed for reasonable combinations of model estimation periods and clinically relevant prediction horizons for CFB. Results: Our models have shown that a clinically relevant time horizon for CFB prediction with the combination of 48 h estimation time and 8-16 h prediction time achieves high accuracy. With an 8-h prediction time, nearly 50% of CFB predictions are within ±0.5 L, and 77% are still within the clinically acceptable range of ±1.0 L. Conclusion: Our study has provided a promising proof of principle and may form the basis for further efforts in the development of computational models for fluid prediction that do not require large datasets for training and validation, as is the case with machine learning or AI-based models. The adaptive transfer function approach allows estimation of CFB course on a dynamically changing patient fluid balance system by simulating the response to the current fluid management regime, providing a useful digital tool for clinicians in daily intensive care.

The Clinical Application of Machine Learning Models for Risk Analysis of Ramp Lesions in Anterior Cruciate Ligament Injuries

BACKGROUND

Peripheral tears of the posterior horn of the medial meniscus, known as "ramp lesions," are commonly found in anterior cruciate ligament (ACL)-deficient knees but are frequently missed on routine evaluation.

PURPOSE

To predict the presence of ramp lesions in ACL-deficient knees using machine learning methods with associated risk factors.

STUDY DESIGN

Cohort study (Diagnosis); Level of evidence, 2.

METHODS

This study included 362 patients who underwent ACL reconstruction between June 2010 and March 2019. The exclusion criteria were combined fractures and multiple ligament injuries, except for medial collateral ligament injuries. Patients were grouped according to the presence of ramp lesions on arthroscopic surgery. Binary logistic regression was used to analyze risk factors including age, sex, body mass index, time from injury to surgery (≥3 or <3 months), mechanism of injury (contact or noncontact), side-to-side laxity, pivot-shift grade, medial and lateral tibial/meniscal slope, location of bone contusion, mechanical axis angle, and lateral femoral condyle (LFC) ratio. The receiver operating characteristic curve and area under the curve were also evaluated.

RESULTS

Ramp lesions were identified in 112 patients (30.9%). The risk for ramp lesions increased with steeper medial tibial and meniscal slopes, higher knee laxity, and an increased LFC ratio. Comparing the final performance of all models, the random forest model yielded the best performance (area under the curve: 0.944), although there were no significant differences among the models (P > .05). The cut-off values for the presence of ramp lesions on receiver operating characteristic analysis were as follows: medial tibial slope >5.5° (P < .001), medial meniscal slope >5.0° (P < .001), and LFC ratio >71.3% (P = .033).

CONCLUSION

Steep medial tibial and meniscal slopes, an increased LFC ratio, and higher knee rotatory laxity were observed risk factors for ramp lesions in patients with an ACL injury. The prediction model of this study could be used as a supplementary diagnostic tool for ramp lesions in ACL-injured knees. In general, care should be taken in patients with ramp lesions and its risk factors during ACL reconstruction.

Clinician Perspectives on Barriers and Enablers to Implementing an Inpatient Oncology Early Warning System: A Mixed-Methods Study

PURPOSE

To elicit end-user and stakeholder perceptions regarding design and implementation of an inpatient clinical deterioration early warning system (EWS) for oncology patients to better fit routine clinical practices and enhance clinical impact.

METHODS

In an explanatory-sequential mixed-methods study, we evaluated a stakeholder-informed oncology early warning system (OncEWS) using surveys and semistructured interviews. Stakeholders were physicians, advanced practice providers (APPs), and nurses. For qualitative data, we used grounded theory and thematic content analysis via the constant comparative method to identify determinants of OncEWS implementation.

RESULTS

Survey respondents generally agreed that an oncology-focused EWS could add value beyond clinical judgment, with nurses endorsing this notion significantly more strongly than other clinicians (nurse: median 5 on a 6-point scale [6 = strongly agree], interquartile range 4-5; doctors/advanced practice providers: 4 [4-5]; P = .005). However, some respondents would not trust an EWS to identify risk accurately (n = 36 [42%] somewhat or very concerned), while others were concerned that institutional culture would not embrace such an EWS (n = 17 [28%]).Interviews highlighted important aspects of the EWS and the local context that might facilitate implementation, including (1) a model tailored to the subtleties of oncology patients, (2) transparent model information, and (3) nursing-centric workflows. Interviewees raised the importance of sepsis as a common and high-risk deterioration syndrome.

CONCLUSION

Stakeholders prioritized maximizing the degree to which the OncEWS is understandable, informative, actionable, and workflow-complementary, and perceived these factors to be key for translation into clinical benefit.

Pediatric Crohn's disease diagnosis aid via genomic analysis and machine learning

Introduction

Determination of pediatric Crohn's disease (CD) remains a major diagnostic challenge. However, the rapidly emerging field of artificial intelligence has demonstrated promise in developing diagnostic models for intractable diseases.

Methods

We propose an artificial neural network model of 8 gene markers identified by 4 classification algorithms based on Gene Expression Omnibus database for diagnostic of pediatric CD.

Results

The model achieved over 85% accuracy and area under ROC curve value in both training set and testing set for diagnosing pediatric CD. Additionally, immune infiltration analysis was performed to address why these markers can be integrated to develop a diagnostic model.

Conclusion

This study supports further clinical facilitation of precise disease diagnosis by integrating genomics and machine learning algorithms in open-access database.

A Predictive Machine Learning Tool for Asthma Exacerbations: Results from a 12-Week, Open-Label Study Using an Electronic Multi-Dose Dry Powder Inhaler with Integrated Sensors

PURPOSE

Machine learning models informed by sensor data inputs have the potential to provide individualized predictions of asthma deterioration. This study aimed to determine if data from an integrated digital inhaler could be used to develop a machine learning model capable of predicting impending exacerbations.

PATIENTS AND METHODS

Adult patients with poorly controlled asthma were enrolled in a 12-week, open-label study using ProAir® Digihaler®, an electronic multi-dose dry powder inhaler (eMDPI) with integrated sensors, as reliever medication (albuterol, 90 µg/dose; 1-2 inhalations every 4 hours, as needed). Throughout the study, the eMDPI recorded inhaler use, peak inspiratory flow (PIF), inhalation volume, inhalation duration, and time to PIF. A model predictive of impending exacerbations was generated by applying machine learning techniques to data downloaded from the inhalers, together with clinical and demographic information. The generated model was evaluated by receiver operating characteristic area under curve (ROC AUC) analysis.

RESULTS

Of 360 patients included in the predictive analysis, 64 experienced a total of 78 exacerbations. Increased albuterol use preceded exacerbations; the mean number of inhalations in the 24-hours preceding an exacerbation was 7.3 (standard deviation 17.3). The machine learning model, using gradient-boosting trees with data from the eMDPI and baseline patient characteristics, predicted an impending exacerbation over the following 5 days with an ROC AUC of 0.83 (95% confidence interval: 0.77-0.90). The feature of the model with the highest weight was the mean number of daily inhalations during the 4 days prior to the day the prediction was made.

CONCLUSION

A machine learning model to predict impending asthma exacerbations using data from the eMDPI was successfully developed. This approach may support a shift from reactive care to proactive, preventative, and personalized management of chronic respiratory diseases.

Machine learning models for predicting in-hospital mortality in patient with sepsis: Analysis of vital sign dynamics

Purpose

To build machine learning models for predicting the risk of in-hospital death in patients with sepsis within 48 h, using only dynamic changes in the patient's vital signs.

Methods

This retrospective observational cohort study enrolled septic patients from five emergency departments (ED) in Taiwan. We adopted seven variables, i.e., age, sex, systolic blood pressure, diastolic blood pressure, heart rate, respiratory rate, and body temperature.

Results

Among all 353,253 visits, after excluding 159,607 visits (45%), the study group consisted of 193,646 ED visits. With a leading time of 6 h, the convolutional neural networks (CNNs), long short-term memory (LSTM), and random forest (RF) had accuracy rates of 0.905, 0.817, and 0.835, respectively, and the area under the receiver operating characteristic curve (AUC) was 0.840, 0.761, and 0.770, respectively. With a leading time of 48 h, the CNN, LSTM, and RF achieved accuracy rates of 0.828, 0759, and 0.805, respectively, and an AUC of 0.811, 0.734, and 0.776, respectively.

Conclusion

By analyzing dynamic vital sign data, machine learning models can predict mortality in septic patients within 6 to 48 h of admission. The performance of the testing models is more accurate if the lead time is closer to the event.

Prediction of septic and hypovolemic shock in intensive care unit patients using machine learning

OBJECTIVE

To create and validate a model for predicting septic or hypovolemic shock from easily obtainable variables collected from patients at admission to an intensive care unit.

METHODS

A predictive modeling study with concurrent cohort data was conducted in a hospital in the interior of northeastern Brazil. Patients aged 18 years or older who were not using vasoactive drugs on the day of admission and were hospitalized from November 2020 to July 2021 were included. The Decision Tree, Random Forest, AdaBoost, Gradient Boosting and XGBoost classification algorithms were tested for use in building the model. The validation method used was k-fold cross validation. The evaluation metrics used were recall, precision and area under the Receiver Operating Characteristic curve.

RESULTS

A total of 720 patients were used to create and validate the model. The models showed high predictive capacity with areas under the Receiver Operating Characteristic curve of 0.979; 0.999; 0.980; 0.998 and 1.00 for the Decision Tree, Random Forest, AdaBoost, Gradient Boosting and XGBoost algorithms, respectively.

CONCLUSION

The predictive model created and validated showed a high ability to predict septic and hypovolemic shock from the time of admission of patients to the intensive care unit.

Machine Learning for Acute Kidney Injury Prediction in the Intensive Care Unit

Machine learning is the field of artificial intelligence in which computers are trained to make predictions or to identify patterns in data through complex mathematical algorithms. It has great potential in critical care to predict outcomes, such as acute kidney injury, and can be used for prognosis and to suggest management strategies. Machine learning can also be used as a research tool to advance our clinical and biochemical understanding of acute kidney injury. In this review, we introduce basic concepts in machine learning and review recent research in each of these domains.

A 29-MRNA HOST RESPONSE WHOLE-BLOOD SIGNATURE IMPROVES PREDICTION OF 28-DAY MORTALITY AND 7-DAY INTENSIVE CARE UNIT CARE IN ADULTS PRESENTING TO THE EMERGENCY DEPARTMENT WITH SUSPECTED ACUTE INFECTION AND/OR SEPSIS

ABSTRACT

Background: Risk stratification of emergency department patients with suspected acute infections and/or suspected sepsis remains challenging. We prospectively validated a 29-messenger RNA host response classifier for predicting severity in these patients. Methods: We enrolled adults presenting with suspected acute infections and at least one vital sign abnormality to six emergency departments in Greece. Twenty-nine target host RNAs were quantified on NanoString nCounter and analyzed with the Inflammatix Severity 2 (IMX-SEV-2) classifier to determine risk scores as low, moderate, and high severity. Performance of IMX-SEV-2 for prediction of 28-day mortality was compared with that of lactate, procalcitonin, and quick sequential organ failure assessment (qSOFA). Results: A total of 397 individuals were enrolled; 38 individuals (9.6%) died within 28 days. Inflammatix Severity 2 classifier predicted 28-day mortality with an area under the receiver operator characteristics curve of 0.82 (95% confidence interval [CI], 0.74-0.90) compared with lactate, 0.66 (95% CI, 0.54-0.77); procalcitonin, 0.67 (95% CI, 0.57-0.78); and qSOFA, 0.81 (95% CI, 0.72-0.89). Combining qSOFA with IMX-SEV-2 improved prognostic accuracy from 0.81 to 0.89 (95% CI, 0.82-0.96). The high-severity (rule-in) interpretation band of IMX-SEV-2 demonstrated 96.9% specificity for predicting 28-day mortality, whereas the low-severity (rule-out) band had a sensitivity of 78.9%. Similarly, IMX-SEV-2 alone accurately predicted the need for day-7 intensive care unit care and further boosted overall accuracy when combined with qSOFA. Conclusions: Inflammatix Severity 2 classifier predicted 28-day mortality and 7-day intensive care unit care with high accuracy and boosted the accuracy of clinical scores when used in combination.

Effectiveness of automated alerting system compared to usual care for the management of sepsis

There is a large body of evidence showing that delayed initiation of sepsis bundle is associated with adverse clinical outcomes in patients with sepsis. However, it is controversial whether electronic automated alerts can help improve clinical outcomes of sepsis. Electronic databases are searched from inception to December 2021 for comparative effectiveness studies comparing automated alerts versus usual care for the management of sepsis. A total of 36 studies are eligible for analysis, including 6 randomized controlled trials and 30 non-randomized studies. There is significant heterogeneity in these studies concerning the study setting, design, and alerting methods. The Bayesian meta-analysis by using pooled effects of non-randomized studies as priors shows a beneficial effect of the alerting system (relative risk [RR]: 0.71; 95% credible interval: 0.62 to 0.81) in reducing mortality. The automated alerting system shows less beneficial effects in the intensive care unit (RR: 0.90; 95% CI: 0.73–1.11) than that in the emergency department (RR: 0.68; 95% CI: 0.51–0.90) and ward (RR: 0.71; 95% CI: 0.61–0.82). Furthermore, machine learning-based prediction methods can reduce mortality by a larger magnitude (RR: 0.56; 95% CI: 0.39–0.80) than rule-based methods (RR: 0.73; 95% CI: 0.63–0.85). The study shows a statistically significant beneficial effect of using the automated alerting system in the management of sepsis. Interestingly, machine learning monitoring systems coupled with better early interventions show promise, especially for patients outside of the intensive care unit.