Identifying acute illness phenotypes via deep temporal interpolation and clustering network on physiologic signatures.

Using clustering analysis for early vital signs, unique patient phenotypes with distinct pathophysiological signatures and clinical outcomes may be revealed and support early clinical decision-making. Phenotyping using early vital signs has proven challenging, as vital signs are typically sampled sporadically. We proposed a novel, deep temporal interpolation and clustering network to simultaneously extract latent representations from irregularly sampled vital signs and derive phenotypes. Four distinct clusters were identified. Phenotype A (18%) had the greatest prevalence of comorbid disease with increased prevalence of prolonged respiratory insufficiency, acute kidney injury, sepsis, and long-term (3-year) mortality. Phenotypes B (33%) and C (31%) had a diffuse pattern of mild organ dysfunction. Phenotype B's favorable short-term clinical outcomes were tempered by the second highest rate of long-term mortality. Phenotype C had favorable clinical outcomes. Phenotype D (17%) exhibited early and persistent hypotension, high incidence of early surgery, and substantial biomarker incidence of inflammation. Despite early and severe illness, phenotype D had the second lowest long-term mortality. After comparing the sequential organ failure assessment scores, the clustering results did not simply provide a recapitulation of previous acuity assessments. This tool may impact triage decisions and have significant implications for clinical decision-support under time constraints and uncertainty.

A Network Analysis of Digital Clock Drawing for Command and Copy Conditions.

Graphomotor and time-based variables from the digital Clock Drawing Test (dCDT) characterize cognitive functions. However, no prior publications have quantified the strength of the associations between digital clock variables as they are produced. We hypothesized that analysis of the production of clock features and their interrelationships, as suggested, will differ between the command and copy test conditions. Older adults aged 65+ completed a digital clock drawing to command and copy conditions. Using a Bayesian hill-climbing algorithm and bootstrapping (10,000 samples), we derived directed acyclic graphs (DAGs) to examine network structure for command and copy dCDT variables. Although the command condition showed moderate associations between variables (µ|ßz|= 0.34) relative to the copy condition (µ|ßz| = 0.25), the copy condition network had more connections (18/18 versus 15/18 command). Network connectivity across command and copy was most influenced by five of the 18 variables. The direction of dependencies followed the order of instructions better in the command condition network. Digitally acquired clock variables relate to one another but differ in network structure when derived from command or copy conditions. Continued analyses of clock drawing production should improve understanding of quintessential normal features to aid in early neurodegenerative disease detection.

FaIRClocks: Fair and Interpretable Representation of the Clock Drawing Test for mitigating classifier bias against lower educational groups.

The clock drawing test (CDT) is a neuropsychological assessment tool to evaluate a patient's cognitive ability. In this study, we developed a Fair and Interpretable Representation of Clock drawing tests (FaIRClocks) to evaluate and mitigate bias against people with lower education while predicting their cognitive status. We represented clock drawings with a 10-dimensional latent embedding using Relevance Factor Variational Autoencoder (RF-VAE) network pretrained on publicly available clock drawings from the National Health and Aging Trends Study (NHATS) dataset. These embeddings were later fine-tuned for predicting three cognitive scores: the Mini-Mental State Examination (MMSE) total score, attention composite z-score (ATT-C), and memory composite z-score (MEM-C). The classifiers were initially tested to see their relative performance in patients with low education (<= 8 years) versus patients with higher education (> 8 years). Results indicated that the initial unweighted classifiers confounded lower education with cognitive impairment, resulting in a 100% type I error rate for this group. Thereby, the samples were re-weighted using multiple fairness metrics to achieve balanced performance. In summary, we report the FaIRClocks model, which a) can identify attention and memory deficits using clock drawings and b) exhibits identical performance between people with higher and lower education levels.

Clinical courses of acute kidney injury in hospitalized patients: a multistate analysis.

Persistence of acute kidney injury (AKI) or insufficient recovery of renal function was associated with reduced long-term survival and life quality. We quantified AKI trajectories and describe transitions through progression and recovery among hospitalized patients. 245,663 encounters from 128,271 patients admitted to UF Health between 2012 and 2019 were retrospectively categorized according to the worst AKI stage experienced within 24-h periods. Multistate models were fit for describing characteristics influencing transitions towards progressed or regressed AKI, discharge, and death. Effects of age, sex, race, admission comorbidities, and prolonged intensive care unit stay (ICU) on transition rates were examined via Cox proportional hazards models. About 20% of encounters had AKI; where 66% of those with AKI had Stage 1 as their worst AKI severity during hospitalization, 18% had Stage 2, and 16% had Stage 3 AKI (12% with kidney replacement therapy (KRT) and 4% without KRT). At 3 days following Stage 1 AKI, 71.1% (70.5-71.6%) were either resolved to No AKI or discharged, while recovery proportion was 38% (37.4-38.6%) and discharge proportion was 7.1% (6.9-7.3%) following AKI Stage 2. At 14 days following Stage 1 AKI, patients with additional frail conditions stay had lower transition proportion towards No AKI or discharge states. Multistate modeling framework is a facilitating mechanism for understanding AKI clinical course and examining characteristics influencing disease process and transition rates.

Machine Learning-Enabled Clinical Information Systems Using Fast Healthcare Interoperability Resources Data Standards: Scoping Review.

Background: Machine learning-enabled clinical information systems (ML-CISs) have the potential to drive health care delivery and research. The Fast Healthcare Interoperability Resources (FHIR) data standard has been increasingly applied in developing these systems. However, methods for applying FHIR to ML-CISs are variable. Objective: This study evaluates and compares the functionalities, strengths, and weaknesses of existing systems and proposes guidelines for optimizing future work with ML-CISs. Methods: Embase, PubMed, and Web of Science were searched for articles describing machine learning systems that were used for clinical data analytics or decision support in compliance with FHIR standards. Information regarding each system's functionality, data sources, formats, security, performance, resource requirements, scalability, strengths, and limitations was compared across systems. Results: A total of 39 articles describing FHIR-based ML-CISs were divided into the following three categories according to their primary focus: clinical decision support systems (n=18), data management and analytic platforms (n=10), or auxiliary modules and application programming interfaces (n=11). Model strengths included novel use of cloud systems, Bayesian networks, visualization strategies, and techniques for translating unstructured or free-text data to FHIR frameworks. Many intelligent systems lacked electronic health record interoperability and externally validated evidence of clinical efficacy. Conclusions: Shortcomings in current ML-CISs can be addressed by incorporating modular and interoperable data management, analytic platforms, secure interinstitutional data exchange, and application programming interfaces with adequate scalability to support both real-time and prospective clinical applications that use electronic health record platforms with diverse implementations.

Digital health and acute kidney injury: consensus report of the 27th Acute Disease Quality Initiative workgroup.

Acute kidney injury (AKI), which is a common complication of acute illnesses, affects the health of individuals in community, acute care and post-acute care settings. Although the recognition, prevention and management of AKI has advanced over the past decades, its incidence and related morbidity, mortality and health care burden remain overwhelming. The rapid growth of digital technologies has provided a new platform to improve patient care, and reports show demonstrable benefits in care processes and, in some instances, in patient outcomes. However, despite great progress, the potential benefits of using digital technology to manage AKI has not yet been fully explored or implemented in clinical practice. Digital health studies in AKI have shown variable evidence of benefits, and the digital divide means that access to digital technologies is not equitable. Upstream research and development costs, limited stakeholder participation and acceptance, and poor scalability of digital health solutions have hindered their widespread implementation and use. Here, we provide recommendations from the Acute Disease Quality Initiative consensus meeting, which involved experts in adult and paediatric nephrology, critical care, pharmacy and data science, at which the use of digital health for risk prediction, prevention, identification and management of AKI and its consequences was discussed.

A deep learning-based dynamic model for predicting acute kidney injury risk severity in postoperative patients.

BACKGROUND: Acute kidney injury is a common postoperative complication affecting between 10% and 30% of surgical patients. Acute kidney injury is associated with increased resource usage and chronic kidney disease development, with more severe acute kidney injury suggesting more aggressive deterioration in clinical outcomes and mortality. METHODS: We considered 42,906 surgical patients admitted to University of Florida Health (n = 51,806) between 2014 and 2021. Acute kidney injury stages were determined using the Kidney Disease Improving Global Outcomes serum creatinine criteria. We developed a recurrent neural network-based model to continuously predict acute kidney injury risk and state in the following 24 hours and compared it with logistic regression, random forest, and multi-layer perceptron models. We used medications, laboratory and vital measurements, and derived features from past one-year records as inputs. We analyzed the proposed model with integrated gradients for enhanced explainability. RESULTS: Postoperative acute kidney injury at any stage developed in 20% (10,664) of the cohort. The recurrent neural network model was more accurate in predicting nearly all categories of next-day acute kidney injury stages (including the no acute kidney injury group). The area under the receiver operating curve and 95% confidence intervals for recurrent neural network and logistic regression models were for no acute kidney injury (0.98 [0.98-0.98] vs 0.93 [0.93-0.93]), stage 1 (0.95 [0.95-0.95] vs. 0.81 [0.80-0.82]), stage 2/3 (0.99 [0.99-0.99] vs 0.96 [0.96-0.97]), and stage 3 with renal replacement therapy (1.0 [1.0-1.0] vs 1.0 [1.0-1.0]. CONCLUSION: The proposed model demonstrates that temporal processing of patient information can lead to more granular and dynamic modeling of acute kidney injury status and result in more continuous and accurate acute kidney injury prediction. We showcase the integrated gradients framework's utility as a mechanism for enhancing model explainability, potentially facilitating clinical trust for future implementation.

Explainable semi-supervised deep learning shows that dementia is associated with small, avocado-shaped clocks with irregularly placed hands.

The clock drawing test is a simple and inexpensive method to screen for cognitive frailties, including dementia. In this study, we used the relevance factor variational autoencoder (RF-VAE), a deep generative neural network, to represent digitized clock drawings from multiple institutions using an optimal number of disentangled latent factors. The model identified unique constructional features of clock drawings in a completely unsupervised manner. These factors were examined by domain experts to be novel and not extensively examined in prior research. The features were informative, as they distinguished dementia from non-dementia patients with an area under receiver operating characteristic (AUC) of 0.86 singly, and 0.96 when combined with participants' demographics. The correlation network of the features depicted the "typical dementia clock" as having a small size, a non-circular or "avocado-like" shape, and incorrectly placed hands. In summary, we report a RF-VAE network whose latent space encoded novel constructional features of clocks that classify dementia from non-dementia patients with high performance.

Artificial intelligence guidance of advanced heart failure therapies: A systematic scoping review.

Introduction: Artificial intelligence can recognize complex patterns in large datasets. It is a promising technology to advance heart failure practice, as many decisions rely on expert opinions in the absence of high-quality data-driven evidence. Methods: We searched Embase, Web of Science, and PubMed databases for articles containing "artificial intelligence," "machine learning," or "deep learning" and any of the phrases "heart transplantation," "ventricular assist device," or "cardiogenic shock" from inception until August 2022. We only included original research addressing post heart transplantation (HTx) or mechanical circulatory support (MCS) clinical care. Review and data extraction were performed in accordance with PRISMA-Scr guidelines. Results: Of 584 unique publications detected, 31 met the inclusion criteria. The majority focused on outcome prediction post HTx (n = 13) and post durable MCS (n = 7), as well as post HTx and MCS management (n = 7, n = 3, respectively). One study addressed temporary mechanical circulatory support. Most studies advocated for rapid integration of AI into clinical practice, acknowledging potential improvements in management guidance and reliability of outcomes prediction. There was a notable paucity of external data validation and integration of multiple data modalities. Conclusion: Our review showed mounting innovation in AI application in management of MCS and HTx, with the largest evidence showing improved mortality outcome prediction.

Clinical Courses of Acute Kidney Injury in Hospitalized Patients: A Multistate Analysis.

OBJECTIVES: We aim to quantify longitudinal acute kidney injury (AKI) trajectories and to describe transitions through progressing and recovery states and outcomes among hospitalized patients using multistate models. METHODS: In this large, longitudinal cohort study, 138,449 adult patients admitted to a quaternary care hospital between 2012 and 2019 were staged based on Kidney Disease: Improving Global Outcomes serum creatinine criteria for the first 14 days of their hospital stay. We fit multistate models to estimate probability of being in a certain clinical state at a given time after entering each one of the AKI stages. We investigated the effects of selected variables on transition rates via Cox proportional hazards regression models. RESULTS: Twenty percent of hospitalized encounters (49,325/246,964) had AKI; among patients with AKI, 66% had Stage 1 AKI, 18% had Stage 2 AKI, and 17% had AKI Stage 3 with or without RRT. At seven days following Stage 1 AKI, 69% (95% confidence interval [CI]: 68.8%-70.5%) were either resolved to No AKI or discharged, while smaller proportions of recovery (26.8%, 95% CI: 26.1%-27.5%) and discharge (17.4%, 95% CI: 16.8%-18.0%) were observed following AKI Stage 2. At 14 days following Stage 1 AKI, patients with more frail conditions (Charlson comorbidity index greater than or equal to 3 and had prolonged ICU stay) had lower proportion of transitioning to No AKI or discharge states. DISCUSSION: Multistate analyses showed that the majority of Stage 2 and higher severity AKI patients could not resolve within seven days; therefore, strategies preventing the persistence or progression of AKI would contribute to the patients' life quality. CONCLUSIONS: We demonstrate multistate modeling framework's utility as a mechanism for a better understanding of the clinical course of AKI with the potential to facilitate treatment and resource planning.

Computable Phenotypes to Characterize Changing Patient Brain Dysfunction in the Intensive Care Unit.

In the United States, more than 5 million patients are admitted annually to ICUs, with ICU mortality of 10%-29% and costs over $82 billion. Acute brain dysfunction status, delirium, is often underdiagnosed or undervalued. This study's objective was to develop automated computable phenotypes for acute brain dysfunction states and describe transitions among brain dysfunction states to illustrate the clinical trajectories of ICU patients. We created two single-center, longitudinal EHR datasets for 48,817 adult patients admitted to an ICU at UFH Gainesville (GNV) and Jacksonville (JAX). We developed algorithms to quantify acute brain dysfunction status including coma, delirium, normal, or death at 12-hour intervals of each ICU admission and to identify acute brain dysfunction phenotypes using continuous acute brain dysfunction status and k-means clustering approach. There were 49,770 admissions for 37,835 patients in UFH GNV dataset and 18,472 admissions for 10,982 patients in UFH JAX dataset. In total, 18% of patients had coma as the worst brain dysfunction status; every 12 hours, around 4%-7% would transit to delirium, 22%-25% would recover, 3%-4% would expire, and 67%-68% would remain in a coma in the ICU. Additionally, 7% of patients had delirium as the worst brain dysfunction status; around 6%-7% would transit to coma, 40%-42% would be no delirium, 1% would expire, and 51%-52% would remain delirium in the ICU. There were three phenotypes: persistent coma/delirium, persistently normal, and transition from coma/delirium to normal almost exclusively in first 48 hours after ICU admission. We developed phenotyping scoring algorithms that determined acute brain dysfunction status every 12 hours while admitted to the ICU. This approach may be useful in developing prognostic and decision-support tools to aid patients and clinicians in decision-making on resource use and escalation of care.

Overtriage, Undertriage, and Value of Care after Major Surgery: An Automated, Explainable Deep Learning-Enabled Classification System.

BACKGROUND: In single-institution studies, overtriaging low-risk postoperative patients to ICUs has been associated with a low value of care; undertriaging high-risk postoperative patients to general wards has been associated with increased mortality and morbidity. This study tested the reproducibility of an automated postoperative triage classification system to generating an actionable, explainable decision support system. STUDY DESIGN: This longitudinal cohort study included adults undergoing inpatient surgery at two university hospitals. Triage classifications were generated by an explainable deep learning model using preoperative and intraoperative electronic health record features. Nearest neighbor algorithms identified risk-matched controls. Primary outcomes were mortality, morbidity, and value of care (inverted risk-adjusted mortality/total direct costs). RESULTS: Among 4,669 ICU admissions, 237 (5.1%) were overtriaged. Compared with 1,021 control ward admissions, overtriaged admissions had similar outcomes but higher costs ($15.9K [interquartile range $9.8K to $22.3K] vs $10.7K [$7.0K to $17.6K], p < 0.001) and lower value of care (0.2 [0.1 to 0.3] vs 1.5 [0.9 to 2.2], p < 0.001). Among 8,594 ward admissions, 1,029 (12.0%) were undertriaged. Compared with 2,498 control ICU admissions, undertriaged admissions had longer hospital length-of-stays (6.4 [3.4 to 12.4] vs 5.4 [2.6 to 10.4] days, p < 0.001); greater incidence of hospital mortality (1.7% vs 0.7%, p = 0.03), cardiac arrest (1.4% vs 0.5%, p = 0.04), and persistent acute kidney injury without renal recovery (5.2% vs 2.8%, p = 0.002); similar costs ($21.8K [$13.3K to $34.9K] vs $21.9K [$13.1K to $36.3K]); and lower value of care (0.8 [0.5 to 1.3] vs 1.2 [0.7 to 2.0], p < 0.001). CONCLUSIONS: Overtriage was associated with low value of care; undertriage was associated with both low value of care and increased mortality and morbidity. The proposed framework for generating automated postoperative triage classifications is reproducible.

Dynamic predictions of postoperative complications from explainable, uncertainty-aware, and multi-task deep neural networks.

Accurate prediction of postoperative complications can inform shared decisions regarding prognosis, preoperative risk-reduction, and postoperative resource use. We hypothesized that multi-task deep learning models would outperform conventional machine learning models in predicting postoperative complications, and that integrating high-resolution intraoperative physiological time series would result in more granular and personalized health representations that would improve prognostication compared to preoperative predictions. In a longitudinal cohort study of 56,242 patients undergoing 67,481 inpatient surgical procedures at a university medical center, we compared deep learning models with random forests and XGBoost for predicting nine common postoperative complications using preoperative, intraoperative, and perioperative patient data. Our study indicated several significant results across experimental settings that suggest the utility of deep learning for capturing more precise representations of patient health for augmented surgical decision support. Multi-task learning improved efficiency by reducing computational resources without compromising predictive performance. Integrated gradients interpretability mechanisms identified potentially modifiable risk factors for each complication. Monte Carlo dropout methods provided a quantitative measure of prediction uncertainty that has the potential to enhance clinical trust. Multi-task learning, interpretability mechanisms, and uncertainty metrics demonstrated potential to facilitate effective clinical implementation.

Building an automated, machine learning-enabled platform for predicting post-operative complications.

Objective. In 2019, the University of Florida College of Medicine launched theMySurgeryRiskalgorithm to predict eight major post-operative complications using automatically extracted data from the electronic health record.Approach. This project was developed in parallel with our Intelligent Critical Care Center and represents a culmination of efforts to build an efficient and accurate model for data processing and predictive analytics.Main Results and Significance. This paper discusses how our model was constructed and improved upon. We highlight the consolidation of the database, processing of fixed and time-series physiologic measurements, development and training of predictive models, and expansion of those models into different aspects of patient assessment and treatment. We end by discussing future directions of the model.

Predictive Modeling for Readmission to Intensive Care: A Systematic Review.

To evaluate the methodologic rigor and predictive performance of models predicting ICU readmission; to understand the characteristics of ideal prediction models; and to elucidate relationships between appropriate triage decisions and patient outcomes. DATA SOURCES: PubMed, Web of Science, Cochrane, and Embase. STUDY SELECTION: Primary literature that reported the development or validation of ICU readmission prediction models within from 2010 to 2021. DATA EXTRACTION: Relevant study information was extracted independently by two authors using the Critical Appraisal and Data Extraction for Systematic Reviews of Prediction Modelling Studies checklist. Bias was evaluated using the Prediction model Risk Of Bias ASsessment Tool. Data sources, modeling methodology, definition of outcomes, performance, and risk of bias were critically evaluated to elucidate relevant relationships. DATA SYNTHESIS: Thirty-three articles describing models were included. Six studies had a high overall risk of bias due to improper inclusion criteria or omission of critical analysis details. Four other studies had an unclear overall risk of bias due to lack of detail describing the analysis. Overall, the most common (50% of studies) source of bias was the filtering of candidate predictors via univariate analysis. The poorest performing models used existing clinical risk or acuity scores such as Acute Physiologic Assessment and Chronic Health Evaluation II, Sequential Organ Failure Assessment, or Stability and Workload Index for Transfer as the sole predictor. The higher-performing ICU readmission prediction models used homogenous patient populations, specifically defined outcomes, and routinely collected predictors that were analyzed over time. CONCLUSIONS: Models predicting ICU readmission can achieve performance advantages by using longitudinal time series modeling, homogenous patient populations, and predictor variables tailored to those populations.

Postoperative Overtriage to an Intensive Care Unit Is Associated With Low Value of Care.

OBJECTIVE: We test the hypothesis that for low-acuity surgical patients, postoperative intensive care unit (ICU) admission is associated with lower value of care compared with ward admission. BACKGROUND: Overtriaging low-acuity patients to ICU consumes valuable resources and may not confer better patient outcomes. Associations among postoperative overtriage, patient outcomes, costs, and value of care have not been previously reported. METHODS: In this longitudinal cohort study, postoperative ICU admissions were classified as overtriaged or appropriately triaged according to machine learning-based patient acuity assessments and requirements for immediate postoperative mechanical ventilation or vasopressor support. The nearest neighbors algorithm identified risk-matched control ward admissions. The primary outcome was value of care, calculated as inverse observed-to-expected mortality ratios divided by total costs. RESULTS: Acuity assessments had an area under the receiver operating characteristic curve of 0.92 in generating predictions for triage classifications. Of 8592 postoperative ICU admissions, 423 (4.9%) were overtriaged. These were matched with 2155 control ward admissions with similar comorbidities, incidence of emergent surgery, immediate postoperative vital signs, and do not resuscitate order placement and rescindment patterns. Compared with controls, overtraiged admissions did not have a lower incidence of any measured complications. Total costs for admission were $16.4K for overtriage and $15.9K for controls ( P =0.03). Value of care was lower for overtriaged admissions [2.9 (2.0-4.0)] compared with controls [24.2 (14.1-34.5), P <0.001]. CONCLUSIONS: Low-acuity postoperative patients who were overtriaged to ICUs had increased total costs, no improvements in outcomes, and received low-value care.

Diurnal Pain Classification in Critically Ill Patients using Machine Learning on Accelerometry and Analgesic Data.

Quantifying pain in patients admitted to intensive care units (ICUs) is challenging due to the increased prevalence of communication barriers in this patient population. Previous research has posited a positive correlation between pain and physical activity in critically ill patients. In this study, we advance this hypothesis by building machine learning classifiers to examine the ability of accelerometer data collected from daily wearables to predict self-reported pain levels experienced by patients in the ICU. We trained multiple Machine Learning (ML) models, including Logistic Regression, CatBoost, and XG-Boost, on statistical features extracted from the accelerometer data combined with previous pain measurements and patient demographics. Following previous studies that showed a change in pain sensitivity in ICU patients at night, we performed the task of pain classification separately for daytime and nighttime pain reports. In the pain versus no-pain classification setting, logistic regression gave the best classifier in daytime (AUC: 0.72, F1-score: 0.72), and CatBoost gave the best classifier at nighttime (AUC: 0.82, F1-score: 0.82). Performance of logistic regression dropped to 0.61 AUC, 0.62 F1-score (mild vs. moderate pain, nighttime), and CatBoost's performance was similarly affected with 0.61 AUC, 0.60 F1-score (moderate vs. severe pain, daytime). The inclusion of analgesic information benefited the classification between moderate and severe pain. SHAP analysis was conducted to find the most significant features in each setting. It assigned the highest importance to accelerometer-related features on all evaluated settings but also showed the contribution of the other features such as age and medications in specific contexts. In conclusion, accelerometer data combined with patient demographics and previous pain measurements can be used to screen painful from painless episodes in the ICU and can be combined with analgesic information to provide moderate classification between painful episodes of different severities.

Dynamic Delirium Prediction in the Intensive Care Unit using Machine Learning on Electronic Health Records.

Delirium is a syndrome of acute brain failure which is prevalent amongst older adults in the Intensive Care Unit (ICU). Incidence of delirium can significantly worsen prognosis and increase mortality, therefore necessitating its rapid and continual assessment in the ICU. Currently, the common approach for delirium assessment is manual and sporadic. Hence, there exists a critical need for a robust and automated system for predicting delirium in the ICU. In this work, we develop a machine learning (ML) system for real-time prediction of delirium using Electronic Health Record (EHR) data. Unlike prior approaches which provide one delirium prediction label per entire ICU stay, our approach provides predictions every 12 hours. We use the latest 12 hours of ICU data, along with patient demographic and medical history data, to predict delirium risk in the next 12-hour window. This enables delirium risk prediction as soon as 12 hours after ICU admission. We train and test four ML classification algorithms on longitudinal EHR data pertaining to 16,327 ICU stays of 13,395 patients covering a total of 56,297 12-hour windows in the ICU to predict the dynamic incidence of delirium. The best performing algorithm was Categorical Boosting which achieved an area under receiver operating characteristic curve (AUROC) of 0.87 (95% Confidence Interval; C.I, 0.86-0.87). The deployment of this ML system in ICUs can enable early identification of delirium, thereby reducing its deleterious impact on long-term adverse outcomes, such as ICU cost, length of stay and mortality.