Prediction model of in-hospital mortality in intensive care unit patients with heart failure: machine learning-based, retrospective analysis of the MIMIC-III database

XGBoost model for predicting erectile dysfunction risk after radical prostatectomy: development and validation using machine learning

BACKGROUND

Erectile dysfunction (ED) is a frequent complication following radical prostatectomy, significantly affecting patients' quality of life. Traditional predictive methods often struggle to capture complex nonlinear risk factors. This study aims to develop a machine learning-based model to improve ED risk stratification and guide personalized management.

METHODS

A total of 1,147 prostate cancer patients were analyzed, among whom 285 (24.85%) developed postoperative ED. Univariate and multivariate analyses identified age, smoking history, Gleason score, prostate volume, T-stage, surgical approach, operative time, intraoperative bleeding, and PCT levels as independent risk factors (P < 0.05). Machine learning models, including XGBoost, Random Forest, Support Vector Machine, and k-Nearest Neighbors, were trained for ED risk prediction. Key predictors included advanced age, smoking history, Gleason score ≥ 8, prostate volume ≥ 40 ml, T-stage, laparoscopic-assisted surgery, and prolonged operative duration.

RESULTS

XGBoost exhibited the highest predictive accuracy (AUC: 0.980 in training; 0.960 in validation), outperforming other models. Calibration curves confirmed strong concordance between predicted and actual probabilities, while decision curve analysis demonstrated superior clinical utility, with XGBoost providing the greatest net benefit. Ten-fold cross-validation indicated stable performance (validation AUC: 0.9127 ± 0.0770; test AUC: 0.9592; accuracy: 0.9111), and external validation confirmed model generalizability (AUC: 0.84). SHAP analysis highlighted key risk contributors, enabling individualized risk assessment and targeted clinical interventions.

CONCLUSION

The XGBoost model exhibited superior predictive performance and clinical applicability in assessing ED risk after radical prostatectomy, offering a robust tool for personalized postoperative management.

Evaluation of machine learning methods for prediction of heart failure mortality and readmission: meta-analysis

BACKGROUND

Heart failure (HF) impacts nearly 6 million individuals in the U.S., with a projected 46% increase by 2030, is creating significant healthcare burdens. Predictive models, particularly machine learning (ML)-based models, offer promising solutions to identify patients at greater risk of adverse outcomes, such as mortality and hospital readmission. This review aims to assess the effectiveness of ML models in predicting HF-related outcomes, with a focus on their potential to improve patient care and clinical decision-making. We aim to assess how effectively machine learning models predict mortality and readmission in heart failure patients to improve clinical outcomes.

METHOD

The study followed PRISMA 2020 guidelines and was registered in the PROSPERO database (CRD42023481167). We conducted a systematic search in PubMed, Scopus, and Web of Science databases using specific keywords related to heart failure, machine learning, mortality and readmission. Extracted data focused on study characteristics, machine learning details, and outcomes, with AUC or c-index used as the primary outcomes for pooling analysis. The PROBAST tool was used to assess bias risk, evaluating models based on participants, predictors, outcomes, and statistical analysis. The meta-analysis pooled AUCs for different machine learning models predicting mortality and readmission. Prediction accuracy data was categorized by timeframes, with high heterogeneity determined by an I² value above 50%, leading to a random-effects model when applicable. Publication bias was assessed using Egger's and Begg's tests, with a p-value below 0.05 considered significant RESULT: A total of 4,505 studies were identified, and after screening, 64 were included in the final analysis, covering 943,941 patients. Of these, 40 studies focused on mortality, 17 on readmission, and 7 on both outcomes. In total, 346 machine learning models were evaluated, with the most common algorithms being random forest, logistic regression, and gradient boosting. The neural network model achieved the highest overall AUC for mortality prediction (0.808), while the support vector machine performed best for readmission prediction (AUC 0.733). The analysis revealed a significant risk of bias, primarily due to reliance on retrospective data and inadequate sample size justification.

CONCLUSION

In conclusion, this review emphasizes the strong potential of ML models in predicting HF readmission and mortality. ML algorithms show promise in improving prognostic accuracy and enabling personalized patient care. However, challenges like model interpretability, generalizability, and clinical integration persist. Overcoming these requires refined ML techniques and a robust regulatory framework to enhance HF outcomes.

Ethical challenges and evolving strategies in the integration of artificial intelligence into clinical practice

Artificial intelligence (AI) has rapidly transformed various sectors, including healthcare, where it holds the potential to transform clinical practice and improve patient outcomes. However, its integration into medical settings brings significant ethical challenges that need careful consideration. This paper examines the current state of AI in healthcare, focusing on five critical ethical concerns: justice and fairness, transparency, patient consent and confidentiality, accountability, and patient-centered and equitable care. These concerns are particularly pressing as AI systems can perpetuate or even exacerbate existing biases, often resulting from non-representative datasets and opaque model development processes. The paper explores how bias, lack of transparency, and challenges in maintaining patient trust can undermine the effectiveness and fairness of AI applications in healthcare. In addition, we review existing frameworks for the regulation and deployment of AI, identifying gaps that limit the widespread adoption of these systems in a just and equitable manner. Our analysis provides recommendations to address these ethical challenges, emphasizing the need for fairness in algorithm design, transparency in model decision-making, and patient-centered approaches to consent and data privacy. By highlighting the importance of continuous ethical scrutiny and collaboration between AI developers, clinicians, and ethicists, we outline pathways for achieving more responsible and inclusive AI implementation in healthcare. These strategies, if adopted, could enhance both the clinical value of AI and the trustworthiness of AI systems among patients and healthcare professionals, ensuring that these technologies serve all populations equitably.

A deep learning model for clinical outcome prediction using longitudinal inpatient electronic health records

Objectives

Recent advances in deep learning show significant potential in analyzing continuous monitoring electronic health records (EHR) data for clinical outcome prediction. We aim to develop a Transformer-based, Encounter-level Clinical Outcome (TECO) model to predict mortality in the intensive care unit (ICU) using inpatient EHR data.

Materials and Methods

The TECO model was developed using multiple baseline and time-dependent clinical variables from 2579 hospitalized COVID-19 patients to predict ICU mortality and was validated externally in an acute respiratory distress syndrome cohort (n = 2799) and a sepsis cohort (n = 6622) from the Medical Information Mart for Intensive Care IV (MIMIC-IV). Model performance was evaluated based on the area under the receiver operating characteristic (AUC) and compared with Epic Deterioration Index (EDI), random forest (RF), and extreme gradient boosting (XGBoost).

Results

In the COVID-19 development dataset, TECO achieved higher AUC (0.89-0.97) across various time intervals compared to EDI (0.86-0.95), RF (0.87-0.96), and XGBoost (0.88-0.96). In the 2 MIMIC testing datasets (EDI not available), TECO yielded higher AUC (0.65-0.77) than RF (0.59-0.75) and XGBoost (0.59-0.74). In addition, TECO was able to identify clinically interpretable features that were correlated with the outcome.

Discussion

The TECO model outperformed proprietary metrics and conventional machine learning models in predicting ICU mortality among patients with COVID-19, widespread inflammation, respiratory illness, and other organ failures.

Conclusion

The TECO model demonstrates a strong capability for predicting ICU mortality using continuous monitoring data. While further validation is needed, TECO has the potential to serve as a powerful early warning tool across various diseases in inpatient settings.

In-Hospital Mortality Prediction among Intensive Care Unit Patients with Acute Ischemic Stroke: A Machine Learning Approach

Background: Acute ischemic stroke is a leading cause of death in the United States. Identifying patients with stroke at high risk of mortality is crucial for timely intervention and optimal resource allocation. This study aims to develop and validate machine learning-based models to predict in-hospital mortality risk for intensive care unit (ICU) patients with acute ischemic stroke and identify important associated factors. Methods: Our data include 3,489 acute ischemic stroke admissions to the ICU for patients not discharged or dead within 48 h from the Medical Information Mart for Intensive Care-IV (MIMIC-IV) database. Demographic, hospitalization type, procedure, medication, intake (intravenous and oral), laboratory, vital signs, and clinical assessment [e.g., Glasgow Coma Scale Scores (GCS)] during the initial 48 h of admissions were used to predict in-hospital mortality after 48 h of ICU admission. We explored 3 machine learning models (random forests, logistic regression, and XGBoost) and applied Bayesian optimization for hyperparameter tuning. Important features were identified using learned coefficients. Results: Experiments show that XGBoost tuned for area under the receiver operating characteristic curve (AUC ROC) was the best performing model (AUC ROC 0.86, F1 0.52), compared to random forests (AUC ROC 0.85, F1 0.47) and logistic regression (AUC ROC 0.75, F1 0.40). Top features include GCS, blood urea nitrogen, and Richmond RASS score. The model also demonstrates good fairness for males versus females and across racial/ethnic groups. Conclusions: Machine learning has shown great potential in predicting in-hospital mortality risk for people with acute ischemic stroke in the ICU setting. However, more ethical considerations need to be applied to ensure that performance differences across different racial/ethnic groups will not exacerbate existing health disparities and will not harm historically marginalized populations.

Machine learning score to predict in-hospital outcomes in patients hospitalized in cardiac intensive care unit

Aims

Although some scores based on traditional statistical methods are available for risk stratification in patients hospitalized in cardiac intensive care units (CICUs), the interest of machine learning (ML) methods for risk stratification in this field is not well established. We aimed to build an ML model to predict in-hospital major adverse events (MAE) in patients hospitalized in CICU.

Methods and results

In April 2021, a French national prospective multicentre study involving 39 centres included all consecutive patients admitted to CICU. The primary outcome was in-hospital MAE, including death, resuscitated cardiac arrest, or cardiogenic shock. Using 31 randomly assigned centres as an index cohort (divided into training and testing sets), several ML models were evaluated to predict in-hospital MAE. The eight remaining centres were used as an external validation cohort. Among 1499 consecutive patients included (aged 64 ± 15 years, 70% male), 67 had in-hospital MAE (4.3%). Out of 28 clinical, biological, ECG, and echocardiographic variables, seven were selected to predict MAE in the training set (n = 844). Boosted cost-sensitive C5.0 technique showed the best performance compared with other ML methods [receiver operating characteristic area under the curve (AUROC) = 0.90, precision-recall AUC = 0.57, F1 score = 0.5]. Our ML score showed a better performance than existing scores (AUROC: ML score = 0.90 vs. Thrombolysis In Myocardial Infarction (TIMI) score: 0.56, Global Registry of Acute Coronary Events (GRACE) score: 0.52, Acute Heart Failure (ACUTE-HF) score: 0.65; all P < 0.05). Machine learning score also showed excellent performance in the external cohort (AUROC = 0.88).

Conclusion

This new ML score is the first to demonstrate improved performance in predicting in-hospital outcomes over existing scores in patients admitted to the intensive care unit based on seven simple and rapid clinical and echocardiographic variables.

Trial Registration

ClinicalTrials.gov Identifier: NCT05063097.

Enhanced heart failure mortality prediction through model-independent hybrid feature selection and explainable machine learning

Heart failure (HF) remains a significant public health challenge with high mortality rates. Machine learning (ML) techniques offer a promising approach to predict HF mortality, potentially improving clinical outcomes. However, the effectiveness of these techniques heavily depends on the quality and relevance of the features used. This study introduces a novel hybrid feature selection methodology that combines Extremely Randomized Trees (Extra-Trees) and non-linear correlation measures to enhance 1-year all-cause mortality prediction in HF patients using echocardiographic and key demographic data. Unlike existing feature selection methods that are often tied to specific ML models and produce inconsistent feature sets across different algorithms, our proposed approach is model-independent, ensuring robustness and generalizability. Moreover, the optimal number of predictive features is identified through loss graph inspection, leading to a compact and highly informative subset of seven features. We trained and evaluated seven widely-used ML models on both the full feature set and the selected subset, finding that most models maintained or improved their predictive performance despite an 80% reduction in features. Model interpretability was enhanced using SHapley Additive exPlanations (SHAP), allowing for a detailed examination of how individual features influence predictions. To further assess its effectiveness, we compared our methodology against widely known feature selection techniques across all seven ML models. The results underscore the superiority of our proposed feature set in accurately predicting HF mortality over conventional methods, offering new opportunities for personalized management strategies based on a streamlined and explainable feature subset.

Prediction of the Risk of Adverse Clinical Outcomes with Machine Learning Techniques in Patients with Noncommunicable Diseases

Decision-making in chronic diseases guided by clinical decision support systems that use models including multiple variables based on artificial intelligence requires scientific validation in different populations to optimize the use of limited human, financial, and clinical resources in healthcare systems worldwide. This cohort study evaluated three machine learning algorithms-XGBoost, Elastic Net logistic regression, and an Artificial Neural Network-to develop a prediction model for three outcomes: mortality, hospitalization, and emergency department visits. The objective was to build a clinical decision support system for patients with noncommunicable diseases treated at the Alma Mater Hospital complex in Medellín, Colombia. We collected 4845 electronic medical record entries from 5000 patients included in the study. The median age was 71.83 years, with 63.8% women and 29.7% receiving home care. The most prevalent medical conditions were diabetes (52.9%), hypertension (67.2%), dyslipidemia (57.3%), and COPD (19.4%). For mortality prediction, the Elastic Net logistic regression model achieved an AUCROC of 0.883 (95% CI: 0.848-0.917), the XGBoost model reached an AUCROC of 0.896 (95% CI: 0.865-0.927), and the Neural Network achieved 0.886 (95% CI: 0.853-0.916). For hospitalization, the Elastic Net model had an AUCROC of 0.952 (95% CI: 0.937-0.965), the XGBoost model achieved 0.963 (95% CI: 0.952-0.974), and the Neural Network scored 0.932 (95% CI: 0.915-0.948). For emergency department visits, the AUCROC values were 0.980 (95% CI: 0.971-0.987) for Elastic Net, 0.977 (95% CI: 0.967-0.986) for XGBoost, and 0.976 (95% CI: 0.968-0.982) for the neural network. A dashboard was developed to interact with an ensemble risk categorization segmenting patient risk in the cohort to aid in clinical decision-making. A clinical decision support system based on artificial intelligence using electronic medical records possibly can help segmenting the risk in populations with Noncommunicable Diseases for effective decision-making.

A deep learning model for clinical outcome prediction using longitudinal inpatient electronic health records

Objective

Recent advances in deep learning show significant potential in analyzing continuous monitoring electronic health records (EHR) data for clinical outcome prediction. We aim to develop a Transformer-based, Encounter-level Clinical Outcome (TECO) model to predict mortality in the intensive care unit (ICU) using inpatient EHR data.

Materials and Methods

TECO was developed using multiple baseline and time-dependent clinical variables from 2579 hospitalized COVID-19 patients to predict ICU mortality, and was validated externally in an ARDS cohort (n=2799) and a sepsis cohort (n=6622) from the Medical Information Mart for Intensive Care (MIMIC)-IV. Model performance was evaluated based on area under the receiver operating characteristic (AUC) and compared with Epic Deterioration Index (EDI), random forest (RF), and extreme gradient boosting (XGBoost).

Results

In the COVID-19 development dataset, TECO achieved higher AUC (0.89-0.97) across various time intervals compared to EDI (0.86-0.95), RF (0.87-0.96), and XGBoost (0.88-0.96). In the two MIMIC testing datasets (EDI not available), TECO yielded higher AUC (0.65-0.76) than RF (0.57-0.73) and XGBoost (0.57-0.73). In addition, TECO was able to identify clinically interpretable features that were correlated with the outcome.

Discussion

TECO outperformed proprietary metrics and conventional machine learning models in predicting ICU mortality among COVID-19 and non-COVID-19 patients.

Conclusions

TECO demonstrates a strong capability for predicting ICU mortality using continuous monitoring data. While further validation is needed, TECO has the potential to serve as a powerful early warning tool across various diseases in inpatient settings.

Prioritizing Patient Selection in Clinical Trials: A Machine Learning Algorithm for Dynamic Prediction of In-Hospital Mortality for ICU Admitted Patients Using Repeated Measurement Data

Background: A machine learning prognostic mortality scoring system was developed to address challenges in patient selection for clinical trials within the Intensive Care Unit (ICU) environment. The algorithm incorporates Red blood cell Distribution Width (RDW) data and other demographic characteristics to predict ICU mortality alongside existing ICU mortality scoring systems like Simplified Acute Physiology Score (SAPS). Methods: The developed algorithm, defined as a Mixed-effects logistic Random Forest for binary data (MixRFb), integrates a Random Forest (RF) classification with a mixed-effects model for binary outcomes, accounting for repeated measurement data. Performance comparisons were conducted with RF and the proposed MixRFb algorithms based solely on SAPS scoring, with additional evaluation using a descriptive receiver operating characteristic curve incorporating RDW's predictive mortality ability. Results: MixRFb, incorporating RDW and other covariates, outperforms the SAPS-based variant, achieving an area under the curve of 0.882 compared to 0.814. Age and RDW were identified as the most significant predictors of ICU mortality, as reported by the variable importance plot analysis. Conclusions: The MixRFb algorithm demonstrates superior efficacy in predicting in-hospital mortality and identifies age and RDW as primary predictors. Implementation of this algorithm could facilitate patient selection for clinical trials, thereby improving trial outcomes and strengthening ethical standards. Future research should focus on enriching algorithm robustness, expanding its applicability across diverse clinical settings and patient demographics, and integrating additional predictive markers to improve patient selection capabilities.

Predictive Value of Machine Learning for the Risk of In-Hospital Death in Patients With Heart Failure: A Systematic Review and Meta-Analysis

BACKGROUND

The efficiency of machine learning (ML) based predictive models in predicting in-hospital mortality for heart failure (HF) patients is a topic of debate. In this context, this study's objective is to conduct a meta-analysis to compare and assess existing prognostic models designed for predicting in-hospital mortality in HF patients.

METHODS

A systematic search of databases was conducted, including PubMed, Embase, Web of Science, and Cochrane Library up to January 2023. To ensure comprehensiveness, we performed an additional search in June 2023. The Prediction Model Risk of Bias Assessment Tool was employed to assess the validity and reliability of ML models.

RESULTS

Our analysis incorporated 28 studies involving a total of 106 predictive models based on 14 different ML techniques. In the training data set, these models showed a combined C-index of 0.781, sensitivity of 0.56, and specificity of 0.94. In the validation data set, the models exhibited a combined C-index of 0.758, sensitivity of 0.57, and specificity of 0.84. Logistic regression (LR) was the most frequently used ML algorithm. LR models in the training set had a combined C-index of 0.795, sensitivity of 0.63, and specificity of 0.85, and these measures for LR models in the validation set were 0.751, 0.66, and 0.79, respectively.

CONCLUSIONS

Our study indicates that although ML is increasingly being leveraged to predict in-hospital mortality for HF patients, the predictive performance remains suboptimal. Although these models have relatively high C-index and specificity, their ability to predict positive events is limited, as indicated by their low sensitivity.

An artificial intelligence application to predict prolonged dependence on mechanical ventilation among patients with critical orthopaedic trauma: an establishment and validation study

BACKGROUND

Prolonged dependence on mechanical ventilation is a common occurrence in clinical ICU patients and presents significant challenges for patient care and resource allocation. Predicting prolonged dependence on mechanical ventilation is crucial for improving patient outcomes, preventing ventilator-associated complications, and guiding targeted clinical interventions. However, specific tools for predicting prolonged mechanical ventilation among ICU patients, particularly those with critical orthopaedic trauma, are currently lacking. The purpose of the study was to establish and validate an artificial intelligence (AI) platform to assess the prolonged dependence on mechanical ventilation among patients with critical orthopaedic trauma.

METHODS

This study analyzed 1400 patients with critical orthopaedic trauma who received mechanical ventilation, and the prolonged dependence on mechanical ventilation was defined as not weaning from mechanical ventilation for ≧ 7 days. Patients were randomly classified into a training cohort and a validation cohort based on the ratio of 8:2. Patients in the training cohort were used to establish models using machine learning techniques, including logistic regression (LR), extreme gradient boosting machine (eXGBM), decision tree (DT), random forest (RF), support vector machine (SVM), and light gradient boosting machine (LightGBM), whereas patients in the validation cohort were used to validate these models. The prediction performance of these models was evaluated using discrimination and calibration. A scoring system was used to comprehensively assess and compare the prediction performance of the models, based on ten evaluation metrics. External validation of the model was performed in 122 patients with critical orthopaedic trauma from a university teaching hospital. Furthermore, the optimal model was deployed as an AI calculator, which was accessible online, to assess the risk of prolonged dependence on mechanical ventilation.

RESULTS

Among the developed models, the eXGBM model had the highest score of 50, followed by the LightGBM model (48) and the RF model (37). In detail, the eXGBM model outperformed other models in terms of recall (0.892), Brier score (0.088), log loss (0.291), and calibration slope (0.999), and the model was the second best in terms of area under the curve value (0.949, 95%: 0.933-0.961), accuracy (0.871), F1 score (0.873), and discrimination slope (0.647). The SHAP revealed that the most important five features were respiratory rate, lower limb fracture, glucose, PaO2, and PaCO2. External validation of the eXGBM model also demonstrated favorable prediction performance, with an AUC value of 0.893 (95%CI: 0.819-0.967). The eXGBM model was successfully deployed as an AI platform, which was at https://prolongedmechanicalventilation-lqsfm6ecky6dpd4ybkvohu.streamlit.app/ . By simply clicking the link and inputting features, users were able to obtain the risk of experiencing prolonged dependence on mechanical ventilation for individuals. Based on the risk of prolonged dependence on mechanical ventilation, patients were stratified into the high-risk or the low-risk groups, and corresponding therapeutic interventions were recommended, accordingly.

CONCLUSIONS

The AI model shows potential as a valuable tool for stratifying patients with a high risk of prolonged dependence on mechanical ventilation. The AI model may offer a promising approach for optimizing patient care and resource allocation in critical care settings.

CLINICAL TRIAL NUMBER

Not applicable.

Fifteen-year experience of direct bridge with venoarterial extracorporeal membrane oxygenation to heart transplantation

Objective

The study objective was to evaluate outcomes of patients directly bridged with venoarterial extracorporeal membrane oxygenation to heart transplantation.

Methods

A single-center retrospective study was performed on 1152 adult patients undergoing isolated cardiac transplantation between January 2007 and December 2021. Among these, patients bridged with an extracorporeal membrane oxygenation to transplantation (extracorporeal membrane oxygenation group, n = 317) were compared with standard cohorts of patients (no extracorporeal membrane oxygenation group, n = 835). A period analysis (Era 1, 2007-2013, vs Era 2, 2014-2021) was performed.

Results

Median duration of extracorporeal membrane oxygenation support before transplantation in the extracorporeal membrane oxygenation group was 8 days. Recipients of extracorporeal membrane oxygenation group were younger, with a better renal function and a shorter time on the waiting list. They were allocated to younger donors, with a longer ischemic time. The extracorporeal membrane oxygenation group and no extracorporeal membrane oxygenation group showed similar 1-year and 9-year survivals: 79.2% versus 79.4%, P = .98, and 56.2% versus 53.9%, P = .59, respectively. Period analysis in the extracorporeal membrane oxygenation group showed improved 1- and 9-year survivals in Era 2 compared with Era 1: 82.7% versus 71.1%, P = .021 and 60.4% versus 50.5%, P = .031, respectively. Era 2 was characterized by a higher rate of patients maintained on extracorporeal membrane oxygenation support after transplantation (92% vs 48%, P < .001), inserted mainly by peripheral cannulation (99.51% vs 57%, P < .001), for a shorter median duration after transplantation (5 vs 6 days, P = .033).

Conclusions

Extracorporeal membrane oxygenation as a direct bridge to heart transplantation shows similar outcomes to standard cohorts of patients. In the extracorporeal membrane oxygenation group, the waiting list time is shorter due to the emergency allocation system, and recipients have no evidence of organ dysfunction at the time of transplantation.

Serum total calcium levels as a non-linear predictor of in-hospital mortality in heart failure patients: insights from a retrospective cohort study

BACKGROUND

Calcium is pivotal in the regulation of bodily homeostasis, with numerous studies highlighting its link to cardiovascular disease in the adult population. However, the relationship between serum calcium levels and the prognosis of heart failure (HF) patients is not clear. This study explored the association between serum total calcium (STC) and in-hospital mortality in patients with HF.

METHODS

Clinical data of 1,176 patients with HF were obtained from the Multiparametric Intelligent Monitoring in Intensive Care III (MIMIC-III) database. The patients were categorized into STC quartiles, and baseline characteristics were comprehensively analyzed. Univariate and multivariate analyses were employed to identify factors associated with in-hospital mortality. To explore the non-linear relationship between STC and mortality, a two-piecewise linear regression model was applied. Subgroup analyses were conducted to identify potential confounding variables.

RESULTS

In this cohort, 159 (13.53%) patients experienced in-hospital mortality. Significant differences in various parameters were observed among STC quartiles. Univariate analysis identified numerous factors associated with mortality. Multivariate analysis confirmed STC as an independent predictor of in-hospital mortality, with a negative association persisting even after adjusting for confounding factors (odds ratio [OR]: 0.49, 95%CI: 0.32-0.76; P = 0.0016). Non-linear analysis revealed an inflection point at 8.41 mg/dL, below which the risk of in-hospital death significantly increased (OR: 0.26, 95%CI: 0.12-0.55; P = 0.0005). Subgroup analyses indicated a pronounced inverse association in patients without atrial fibrillation or chronic obstructive pulmonary disease, as well as those with a left ventricular ejection fraction ≤ 50%.

CONCLUSION

This study identified STC as an independent predictor of in-hospital mortality in HF patients, with a non-linear relationship.

Machine learning for predicting in-hospital mortality in elderly patients with heart failure combined with hypertension: a multicenter retrospective study

BACKGROUND

Heart failure combined with hypertension is a major contributor for elderly patients (≥ 65 years) to in-hospital mortality. However, there are very few models to predict in-hospital mortality in such elderly patients. We aimed to develop and test an individualized machine learning model to assess risk factors and predict in-hospital mortality in in these patients.

METHODS

From January 2012 to December 2021, this study collected data on elderly patients with heart failure and hypertension from the Chongqing Medical University Medical Data Platform. Least absolute shrinkage and the selection operator was used for recognizing key clinical variables. The optimal predictive model was chosen among eight machine learning algorithms on the basis of area under curve. SHapley Additive exPlanations and Local Interpretable Model-agnostic Explanations was employed to interpret the outcome of the predictive model.

RESULTS

This study ultimately comprised 4647 elderly individuals with hypertension and heart failure. The Random Forest model was chosen with the highest area under curve for 0.850 (95% CI 0.789-0.897), high accuracy for 0.738, recall 0.837, specificity 0.734 and brier score 0.178. According to SHapley Additive exPlanations results, the most related factors for in-hospital mortality in elderly patients with heart failure and hypertension were urea, length of stay, neutrophils, albumin and high-density lipoprotein cholesterol.

CONCLUSIONS

This study developed eight machine learning models to predict in-hospital mortality in elderly patients with hypertension as well as heart failure. Compared to other algorithms, the Random Forest model performed significantly better. Our study successfully predicted in-hospital mortality and identified the factors most associated with in-hospital mortality.

Predictive Analytics in Heart Failure Risk, Readmission, and Mortality Prediction: A Review

Heart failure is a leading cause of death among people worldwide. The cost of treatment can be prohibitive, and early prediction of heart failure would reduce treatment costs to patients and hospitals. Improved readmission prediction would also greatly help hospitals, allowing them to manage their treatment programs and budgets better. This literature review aims to summarize recent studies of predictive analytics models that have been constructed to predict heart failure risk, readmission, and mortality. Random forest, logistic regression, neural networks, and XGBoost were among the most common modeling techniques applied. Most selected studies leveraged structured electronic health record data, including demographics, clinical values, lifestyle, and comorbidities, with some incorporating unstructured clinical notes. Preprocessing through imputation and feature selection were frequently employed in building the predictive analytics models. The reviewed studies exhibit demonstrated promise for predictive analytics in improving early heart failure diagnosis, readmission risk stratification, and mortality prediction. This review study highlights rising research activities and the potential of predictive analytics, especially the implementation of machine learning, in advancing heart failure outcomes. Further rigorous, comprehensive syntheses and head-to-head benchmarking of predictive models are needed to derive robust evidence for clinical adoption.

Total bilirubin level is associated with acute kidney injury in neonates admitted to the neonatal intensive care units: based on MIMIC-III database

OBJECTIVE

The objective of this study was to investigate the association between total bilirubin and acute kidney injury (AKI) in neonates admitted to neonatal intensive care units (NICU).

METHODS

All data utilized were extracted from Medical Information Mart for Intensive Care-III (MIMIC-III) in this retrospective cohort study. The primary outcome was the occurrence of AKI during hospitalization in the NICU, and the exposure was the initial measurement of total bilirubin levels within 24 h of neonatal admission to the NICU. The relationship between serum total bilirubin and AKI was evaluated by employing univariate and multivariate logistic regression models. Additionally, subgroup analyses were conducted based on birth weight, sepsis, and mechanical ventilation.

RESULTS

This retrospective cohort study included a population of 1,726 neonates, and 95 neonates developed AKI. Total bilirubin, as a continuous variable, was linked with decreased AKI risk among neonates admitted to the NICU [odds ratio (OR) = 0.77, 95% confidence interval (CI): 0.64-0.92]. Similarly, when total bilirubin levels were categorized by tertiles, tertiles 3 showed a significant association with decreased AKI risk (OR = 0.39, 95%CI: 0.19-0.83). The relationship of total bilirubin level and AKI was also existent among neonates admitted to the NICU who were underweight, had not sepsis, and received mechanical ventilation.

CONCLUSION

Total bilirubin level may be a protective factor for the risk of developing AKI.

Advancing ICU patient care with a Real-Time predictive model for mechanical Power to mitigate VILI

BACKGROUND

Invasive Mechanical Ventilation (IMV) in Intensive Care Units (ICU) significantly increases the risk of Ventilator-Induced Lung Injury (VILI), necessitating careful management of mechanical power (MP). This study aims to develop a real-time predictive model of MP utilizing Artificial Intelligence to mitigate VILI.

METHODOLOGY

A retrospective observational study was conducted, extracting patient data from Clinical Information Systems from 2018 to 2022. Patients over 18 years old with more than 6 h of IMV were selected. Continuous data on IMV variables, laboratory data, monitoring, procedures, demographic data, type of admission, reason for admission, and APACHE II at admission were extracted. The variables with the highest correlation to MP were used for prediction and IMV data was grouped in 15-minute intervals using the mean. A mixed neural network model was developed to forecast MP 15 min in advance, using IMV data from 6 h before the prediction and current patient status. The model's ability to predict future MP was analyzed and compared to a baseline model predicting the future value of MP as equal to the current value.

RESULTS

The cohort consisted of 1967 patients after applying inclusion criteria, with a median age of 63 years and 66.9 % male. The deep learning model achieved a mean squared error of 2.79 in the test set, indicating a 20 % improvement over the baseline model. It demonstrated high accuracy (94 %) in predicting whether MP would exceed a critical threshold of 18 J/min, which correlates with increased mortality. The integration of this model into a web platform allows clinicians real-time access to MP predictions, facilitating timely adjustments to ventilation settings.

CONCLUSIONS

The study successfully developed and integrated in clinical practice a predictive model for MP. This model will assist clinicians allowing for the adjustment of ventilatory parameters before lung damage occurs.

Development and validation of novel interpretable survival prediction models based on drug exposures for severe heart failure during vulnerable period

BACKGROUND

Severe heart failure (HF) has a higher mortality during vulnerable period while targeted predictive tools, especially based on drug exposures, to accurately assess its prognoses remain largely unexplored. Therefore, this study aimed to utilize drug information as the main predictor to develop and validate survival models for severe HF patients during this period.

METHODS

We extracted severe HF patients from the MIMIC-IV database (as training and internal validation cohorts) as well as from the MIMIC-III database and local hospital (as external validation cohorts). Three algorithms, including Cox proportional hazards model (CoxPH), random survival forest (RSF), and deep learning survival prediction (DeepSurv), were applied to incorporate the parameters (partial hospitalization information and exposure durations of drugs) for constructing survival prediction models. The model performance was assessed mainly using area under the receiver operator characteristic curve (AUC), brier score (BS), and decision curve analysis (DCA). The model interpretability was determined by the permutation importance and Shapley additive explanations values.

RESULTS

A total of 11,590 patients were included in this study. Among the 3 models, the CoxPH model ultimately included 10 variables, while RSF and DeepSurv models incorporated 24 variables, respectively. All of the 3 models achieved respectable performance metrics while the DeepSurv model exhibited the highest AUC values and relatively lower BS among these models. The DCA also verified that the DeepSurv model had the best clinical practicality.

CONCLUSIONS

The survival prediction tools established in this study can be applied to severe HF patients during vulnerable period by mainly inputting drug treatment duration, thus contributing to optimal clinical decisions prospectively.

Identification and Analysis of Immune Microenvironment-Related Genes for Keloid Risk Prediction and Their Effects on Keloid Proliferation and Migration

Keloid is a kind of proliferative scar with continuous growth, no restriction and easy recurrence, which cannot be cured and bring serious physical injury and psychological burden to patients. The main reason is that the pathological mechanism is not clear. Therefore, this project is expected to reveal the immune microenvironment-related genes and their functions in keloid progression, and provide effective targets for the treatment of keloid. Firstly, 8 kinds of immune infiltrating cells and 19 potential characteristic genes were identified by immune infiltration analysis, ssGSEA, LASSO regression (glmnet algorithm and lars algorithm) and WGCNA, indicating that keloid was closely related to the changes of immune microenvironment. Then, 4 pathological biomarkers of keloid (MAPK1, PTPRC, STAT3 and IL1R1) were identified by differentially analysis, univariate analysis, LASSO regression (lars algorithm), support vector machine recursive feature elimination (SVM-REF) algorithm, multivariate logical regression analysis and six machine learning algorithms. Based on the 4 feature genes, the risk prediction model and nomogram were constructed. Calibration curve and ROC analysis (AUC = 0.930) showed that the model had reliable clinical value. Subsequently, consistent cluster analysis was used to find that there were 2 immune microenvironment subsets in keloid patients, of which subgroup II was immune subgroup. Multiple independent datasets and RT-qPCR showed that the expression trend of the 4 genes was consistent with the analysis. Cell gain-loss experiment confirmed that 4 genes regulated the proliferation and migration of keloid cells. The above data shows that MAPK1, PTPRC, STAT3 and IL1R1 may be personalized therapeutic targets for keloid patients.

A two-tier feature selection method for predicting mortality risk in ICU patients with acute kidney injury

Acute kidney injury (AKI) is one of the most important lethal factors for patients admitted to intensive care units (ICUs), and timely high-risk prognostic assessment and intervention are essential to improving patient prognosis. In this study, a stacking model using the MIMIC-III dataset with a two-tier feature selection approach was developed to predict the risk of in-hospital mortality in ICU patients admitted for AKI. External validation was performed using separate MIMIC-IV and eICU-CRD. The area under the curve (AUC) was calculated using the stacking model, and features were selected using the Boruta and XGBoost feature selection methods. This study compares the performance of a stacking model using two-tier feature selection with a model using single-tier feature selection (XGBoost: 85; Boruta: 83; two-tier: 0.91). The predictive effectiveness of the stacking model was further validated by using different datasets (Validation 1: 0.83; Validation 2: 0.85) and comparing it with a simpler model and traditional clinical scores (SOFA: 0.65; APACH IV: 0.61). In addition, this study combined interpretable techniques and causal inference to analyze the causal relationship between features and predicted outcomes.

Learning Physiological Mechanisms that Predict Adverse Cardiovascular Events in Intensive Care Patients with Chronic Heart Disease

Chronic heart disease is a burdensome, complex, and fatal condition. Learning the mechanisms driving the development of heart disease is key to early risk assessment and intervention. However, many current machine learning approaches lack sufficient interpretability. Using 2,737 patients with chronic heart disease from the MIMIC-III database, we trained an interpretable Tropical Geometry Fuzzy Neural Network to predict one-year occurrence of a severe cardiac procedure or mortality (AUROC=0.663). We present the 20 learned rules which explain the model predictions. We find that the rules are clinically valid and indicate underlying pathologies. We anticipate that with additional development and validation, these rules will aid clinicians in providing preventative care for chronic heart disease patients in intensive care units.

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

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

Improving the Prognostic Evaluation Precision of Hospital Outcomes for Heart Failure Using Admission Notes and Clinical Tabular Data: Multimodal Deep Learning Model

BACKGROUND

Clinical notes contain contextualized information beyond structured data related to patients' past and current health status.

OBJECTIVE

This study aimed to design a multimodal deep learning approach to improve the evaluation precision of hospital outcomes for heart failure (HF) using admission clinical notes and easily collected tabular data.

METHODS

Data for the development and validation of the multimodal model were retrospectively derived from 3 open-access US databases, including the Medical Information Mart for Intensive Care III v1.4 (MIMIC-III) and MIMIC-IV v1.0, collected from a teaching hospital from 2001 to 2019, and the eICU Collaborative Research Database v1.2, collected from 208 hospitals from 2014 to 2015. The study cohorts consisted of all patients with critical HF. The clinical notes, including chief complaint, history of present illness, physical examination, medical history, and admission medication, as well as clinical variables recorded in electronic health records, were analyzed. We developed a deep learning mortality prediction model for in-hospital patients, which underwent complete internal, prospective, and external evaluation. The Integrated Gradients and SHapley Additive exPlanations (SHAP) methods were used to analyze the importance of risk factors.

RESULTS

The study included 9989 (16.4%) patients in the development set, 2497 (14.1%) patients in the internal validation set, 1896 (18.3%) in the prospective validation set, and 7432 (15%) patients in the external validation set. The area under the receiver operating characteristic curve of the models was 0.838 (95% CI 0.827-0.851), 0.849 (95% CI 0.841-0.856), and 0.767 (95% CI 0.762-0.772), for the internal, prospective, and external validation sets, respectively. The area under the receiver operating characteristic curve of the multimodal model outperformed that of the unimodal models in all test sets, and tabular data contributed to higher discrimination. The medical history and physical examination were more useful than other factors in early assessments.

CONCLUSIONS

The multimodal deep learning model for combining admission notes and clinical tabular data showed promising efficacy as a potentially novel method in evaluating the risk of mortality in patients with HF, providing more accurate and timely decision support.

Lactate dehydrogenase to albumin ratio is associated with in-hospital mortality in patients with acute heart failure: Data from the MIMIC-III database

The effect of the lactate dehydrogenase to albumin ratio (LAR) on the survival of patients with acute heart failure (AHF) is unclear. We aimed to analyze the impact of LAR on survival in patients with AHF. We retrieved eligible patients for our study from the Monitoring in Intensive Care Database III. For each patient in our study, we gathered clinical data and demographic information. We conducted multivariate logistic regression modeling and smooth curve fitting to assess whether the LAR score could be used as an independent indicator for predicting the prognosis of AHF patients. A total of 2,177 patients were extracted from the database. Survivors had an average age of 69.88, whereas nonsurvivors had an average age of 71.95. The survivor group had a mean LAR ratio of 13.44, and the nonsurvivor group had a value of 17.38. LAR and in-hospital mortality had a nearly linear correlation, according to smooth curve fitting (P < 0.001). According to multivariate logistic regression, the LAR may be an independent risk factor in predicting the prognosis of patients with AHF (odd ratio = 1.09; P < 0.001). The LAR ratio is an independent risk factor associated with increased in-hospital mortality rates in patients with AHF.

A machine learning-based nomogram model for predicting the recurrence of cystitis glandularis

Background

Cystitis glandularis is a chronic inflammatory disease of the urinary system characterized by high recurrence rates, the reasons for which are still unknown.

Objectives

This study aims to identify potential factors contributing to recurrence and propose a simple and feasible prognostic model through nomogram construction.

Design

Patients with confirmed recurrence based on outpatient visits or readmissions were included in this study, which was subsequently divided into training and validation cohorts.

Methods

Machine learning techniques were utilized to screen for the most important predictors, and these were then employed to construct the nomogram. The reliability of the nomogram was assessed through receiver operating characteristic curve analysis, decision curve analysis, and calibration curves.

Results

A total of 252 patients met the screening criteria and were enrolled in this study. Over the 12-month follow-up period, the relapse rate was found to be 57.14% (n = 144). The five final predictors identified through machine learning were urinary infections, urinary calculi, eosinophil count, lymphocyte count, and serum magnesium. The area under curve values for all three time points assessing recurrence exceeded 0.75. Furthermore, both calibration curves and decision curve analyses indicated good performance of the nomogram.

Conclusion

We have developed a reliable machine learning-based nomogram for predicting recurrence in cystitis glandularis.

Prediction of HER2 status via random forest in 3257 Chinese patients with gastric cancer

The accurate evaluation of human epidermal growth factor receptor 2 (HER2) is crucial for successful trastuzumab-based therapy in individuals with gastric cancer (GC). The present study, involving a retrospective cohort (N = 2865) from Wuhan Union Hospital and a prospective cohort (N = 392) from Renmin Hospital of Wuhan University, evaluated the benefits of clinical features using random forest and logistic regression models for the detection of HER2 status in patients with GC. Patients from the Union cohort were randomly assigned to either a training (N = 2005) or an internal validation (N = 860) group. Data processing and feature selection were done in Python, which was also used to build random forest and logistic regression models for the prediction of HER2 overexpression. The Renmin cohort (N = 392) was used as the external validation group. Ten features were closely correlated with HER2 overexpression, including age, albumin/globulin ratio, globulin, activated partial thromboplastin time, tumor stage, node stage, tumor node metastasis stage, tumor size, tumor differentiation, and neuron-specific enolase (NSE). Random forest and logistic regression had areas under the curve (AUC) of 0.9995 and 0.6653 in the training group and 0.923 and 0.667 in the internal validation group, respectively. When the two predictive models were validated using data from the Renmin cohort, random forest and logistic regression had AUCs of 0.9994 and 0.627, respectively. This is the first multicenter study to predict HER2 overexpression in individuals with GC, based on clinical variables. The random forest model significantly outperformed the logistic regression model.

Predicting 1-, 3-, 5-, and 8-year all-cause mortality in a community-dwelling older adult cohort: relevance for predictive, preventive, and personalized medicine

BACKGROUND

Population aging is a global public health issue involving increased prevalence of age-related diseases, and concomitant burden on medical resources and the economy. Ninety-two diseases have been identified as age-related, accounting for 51.3% of the global adult disease burden. The economic cost per capita for older people over 60 years is 10 times that of the younger population. From the aspects of predictive, preventive, and personalized medicine (PPPM), developing a risk-prediction model can help identify individuals at high risk for all-cause mortality and provide an opportunity for targeted prevention through personalized intervention at an early stage. However, there is still a lack of predictive models to help community-dwelling older adults do well in healthcare.

OBJECTIVES

This study aims to develop an accurate 1-, 3-, 5-, and 8-year all-cause mortality risk-prediction model by using clinical multidimensional variables, and investigate risk factors for 1-, 3-, 5-, and 8-year all-cause mortality in community-dwelling older adults to guide primary prevention.

METHODS

This is a two-center cohort study. Inclusion criteria: (1) community-dwelling adult, (2) resided in the districts of Chaonan or Haojiang for more than 6 months in the past 12 months, and (3) completed a health examination. Exclusion criteria: (1) age less than 60 years, (2) more than 30 incomplete variables, (3) no signed informed consent. The primary outcome of the study was all-cause mortality obtained from face-to-face interviews, telephone interviews, and the medical death database from 2012 to 2021. Finally, we enrolled 5085 community-dwelling adults, 60 years and older, who underwent routine health screening in the Chaonan and Haojiang districts, southern China, from 2012 to 2021. Of them, 3091 participants from Chaonan were recruited as the primary training and internal validation study cohort, while 1994 participants from Haojiang were recruited as the external validation cohort. A total of 95 clinical multidimensional variables, including demographics, lifestyle behaviors, symptoms, medical history, family history, physical examination, laboratory tests, and electrocardiogram (ECG) data were collected to identify candidate risk factors and characteristics. Risk factors were identified using least absolute shrinkage and selection operator (LASSO) models and multivariable Cox proportional hazards regression analysis. A nomogram predictive model for 1-, 3-, 5- and 8-year all-cause mortality was constructed. The accuracy and calibration of the nomogram prediction model were assessed using the concordance index (C-index), integrated Brier score (IBS), receiver operating characteristic (ROC), and calibration curves. The clinical validity of the model was assessed using decision curve analysis (DCA).

RESULTS

Nine independent risk factors for 1-, 3-, 5-, and 8-year all-cause mortality were identified, including increased age, male, alcohol status, higher daily liquor consumption, history of cancer, elevated fasting glucose, lower hemoglobin, higher heart rate, and the occurrence of heart block. The acquisition of risk factor criteria is low cost, easily obtained, convenient for clinical application, and provides new insights and targets for the development of personalized prevention and interventions for high-risk individuals. The areas under the curve (AUC) of the nomogram model were 0.767, 0.776, and 0.806, and the C-indexes were 0.765, 0.775, and 0.797, in the training, internal validation, and external validation sets, respectively. The IBS was less than 0.25, which indicates good calibration. Calibration and decision curves showed that the predicted probabilities were in good agreement with the actual probabilities and had good clinical predictive value for PPPM.

CONCLUSION

The personalized risk prediction model can identify individuals at high risk of all-cause mortality, help offer primary care to prevent all-cause mortality, and provide personalized medical treatment for these high-risk individuals from the PPPM perspective. Strict control of daily liquor consumption, lowering fasting glucose, raising hemoglobin, controlling heart rate, and treatment of heart block could be beneficial for improving survival in elderly populations.

SUPPLEMENTARY INFORMATION

The online version contains supplementary material available at 10.1007/s13167-023-00342-4.

Machine Learning Algorithms Predict Successful Weaning From Mechanical Ventilation Before Intubation: Retrospective Analysis From the Medical Information Mart for Intensive Care IV Database

BACKGROUND

The prediction of successful weaning from mechanical ventilation (MV) in advance of intubation can facilitate discussions regarding end-of-life care before unnecessary intubation.

OBJECTIVE

We aimed to develop a machine learning-based model that predicts successful weaning from ventilator support based on routine clinical and laboratory data taken before or immediately after intubation.

METHODS

We used the Medical Information Mart for Intensive Care IV database, which is an open-access database covering 524,740 admissions of 382,278 patients in Beth Israel Deaconess Medical Center, United States, from 2008 to 2019. We selected adult patients who underwent MV in the intensive care unit (ICU). Clinical and laboratory variables that are considered relevant to the prognosis of the patient in the ICU were selected. Data collected before or within 24 hours of intubation were used to develop machine learning models that predict the probability of successful weaning within 14 days of ventilator support. Developed models were integrated into an ensemble model. Performance metrics were calculated by 5-fold cross-validation for each model, and a permutation feature importance and Shapley additive explanations analysis was conducted to better understand the impacts of individual variables on outcome prediction.

RESULTS

Of the 23,242 patients, 19,025 (81.9%) patients were successfully weaned from MV within 14 days. Using the preselected 46 clinical and laboratory variables, the area under the receiver operating characteristic curve of CatBoost classifier, random forest classifier, and regularized logistic regression classifier models were 0.860 (95% CI 0.852-0.868), 0.855 (95% CI 0.848-0.863), and 0.823 (95% CI 0.813-0.832), respectively. Using the ensemble voting classifier using the 3 models above, the final model revealed the area under the receiver operating characteristic curve of 0.861 (95% CI 0.853-0.869), which was significantly better than that of Simplified Acute Physiology Score II (0.749, 95% CI 0.742-0.756) and Sequential Organ Failure Assessment (0.588, 95% CI 0.566-0.609). The top features included lactate and anion gap. The model's performance achieved a plateau with approximately the top 21 variables.

CONCLUSIONS

We developed machine learning algorithms that can predict successful weaning from MV in advance to intubation in the ICU. Our models can aid the appropriate management for patients who hesitate to decide on ventilator support or meaningless end-of-life care.

Machine learning based readmission and mortality prediction in heart failure patients

This study intends to predict in-hospital and 6-month mortality, as well as 30-day and 90-day hospital readmission, using Machine Learning (ML) approach via conventional features. A total of 737 patients remained after applying the exclusion criteria to 1101 heart failure patients. Thirty-four conventional features were collected for each patient. First, the data were divided into train and test cohorts with a 70-30% ratio. Then train data were normalized using the Z-score method, and its mean and standard deviation were applied to the test data. Subsequently, Boruta, RFE, and MRMR feature selection methods were utilized to select more important features in the training set. In the next step, eight ML approaches were used for modeling. Next, hyperparameters were optimized using tenfold cross-validation and grid search in the train dataset. All model development steps (normalization, feature selection, and hyperparameter optimization) were performed on a train set without touching the hold-out test set. Then, bootstrapping was done 1000 times on the hold-out test data. Finally, the obtained results were evaluated using four metrics: area under the ROC curve (AUC), accuracy (ACC), specificity (SPE), and sensitivity (SEN). The RFE-LR (AUC: 0.91, ACC: 0.84, SPE: 0.84, SEN: 0.83) and Boruta-LR (AUC: 0.90, ACC: 0.85, SPE: 0.85, SEN: 0.83) models generated the best results in terms of in-hospital mortality. In terms of 30-day rehospitalization, Boruta-SVM (AUC: 0.73, ACC: 0.81, SPE: 0.85, SEN: 0.50) and MRMR-LR (AUC: 0.71, ACC: 0.68, SPE: 0.69, SEN: 0.63) models performed the best. The best model for 3-month rehospitalization was MRMR-KNN (AUC: 0.60, ACC: 0.63, SPE: 0.66, SEN: 0.53) and regarding 6-month mortality, the MRMR-LR (AUC: 0.61, ACC: 0.63, SPE: 0.44, SEN: 0.66) and MRMR-NB (AUC: 0.59, ACC: 0.61, SPE: 0.48, SEN: 0.63) models outperformed the others. Reliable models were developed in 30-day rehospitalization and in-hospital mortality using conventional features and ML techniques. Such models can effectively personalize treatment, decision-making, and wiser budget allocation. Obtained results in 3-month rehospitalization and 6-month mortality endpoints were not astonishing and further experiments with additional information are needed to fetch promising results in these endpoints.

Development and validation of a prediction model for in-hospital death in patients with heart failure and atrial fibrillation

BACKGROUND

To develop a prediction model for in-hospital mortality of patients with heart failure (HF) and atrial fibrillation (AF).

METHODS

This cohort study extracted the data of 10,236 patients with HF and AF upon intensive care unit (ICU) from the Medical Information Mart for Intensive Care (MIMIC). The subjects from MIMIC-IV were divided into the training set to construct the prediction model, and the testing set to verify the performance of the model. The samples from MIMIC-III database and eICU-CRD were included as the internal and external validation set to further validate the predictive value of the model, respectively. Univariate and multivariable Logistic regression analyses were used to explore predictors for in-hospital death in patients with HF and AF. The receiver operator characteristic (ROC), calibration curves and the decision curve analysis (DCA) curves were plotted to evaluate the predictive values of the model.

RESULTS

The mean survival time of participants from MIMIC-III was 11.29 ± 10.05 days and the mean survival time of participants from MIMIC-IV was 10.56 ± 9.19 days. Simplified acute physiology score (SAPSII), red blood cell distribution width (RDW), beta-blocker, race, respiratory rate, urine output, coronary artery bypass grafting (CABG), Charlson comorbidity index, renal replacement therapies (RRT), antiarrhythmic, age, and anticoagulation were predictors finally included in the prediction model. The AUC of our prediction model was 0.810 (95%CI: 0.791-0.828) in the training set, 0.757 (95%CI: 0.729-0.786) in the testing set, 0.792 (95%CI: 0.774-0.810) in the internal validation set, and 0.724 (95%CI: 0.687-0.762) in the external validation set. The calibration curves of revealed that the predictive probabilities of our model for the in-hospital death in patients with HF and AF deviated slightly from the ideal model. The DCA curves revealed that the use of our prediction model increased the net benefit than use no model.

CONCLUSION

The prediction model had good discriminative ability, and might provide a tool to timely identify patients with HF complicated with AF who were at high risk of in-hospital mortality.

A novel web-based calculator to predict 30-day all-cause in-hospital mortality for 7,202 elderly patients with heart failure in ICUs: a multicenter retrospective cohort study in the United States

Background and aims

Heart failure (HF) is a significant cause of in-hospital mortality, especially for the elderly admitted to intensive care units (ICUs). This study aimed to develop a web-based calculator to predict 30-day in-hospital mortality for elderly patients with HF in the ICU and found a relationship between risk factors and the predicted probability of death.

Methods and results

Data (N = 4450) from the MIMIC-III/IV database were used for model training and internal testing. Data (N = 2,752) from the eICU-CRD database were used for external validation. The Brier score and area under the curve (AUC) were employed for the assessment of the proposed nomogram. Restrictive cubic splines (RCSs) found the cutoff values of variables. The smooth curve showed the relationship between the variables and the predicted probability of death. A total of 7,202 elderly patients with HF were included in the study, of which 1,212 died. Multivariate logistic regression analysis showed that 30-day mortality of HF patients in ICU was significantly associated with heart rate (HR), 24-h urine output (24h UOP), serum calcium, blood urea nitrogen (BUN), NT-proBNP, SpO2, systolic blood pressure (SBP), and temperature (P < 0.01). The AUC and Brier score of the nomogram were 0.71 (0.67, 0.75) and 0.12 (0.11, 0.15) in the testing set and 0.73 (0.70, 0.75), 0.13 (0.12, 0.15), 0.65 (0.62, 0.68), and 0.13 (0.12, 0.13) in the external validation set, respectively. The RCS plot showed that the cutoff values of variables were HR of 96 bmp, 24h UOP of 1.2 L, serum calcium of 8.7 mg/dL, BUN of 30 mg/dL, NT-pro-BNP of 5121 pg/mL, SpO2 of 93%, SBP of 137 mmHg, and a temperature of 36.4°C.

Conclusion

Decreased temperature, decreased SpO2, decreased 24h UOP, increased NT-proBNP, increased serum BUN, increased or decreased SBP, fast HR, and increased or decreased serum calcium increase the predicted probability of death. The web-based nomogram developed in this study showed good performance in predicting 30-day in-hospital mortality for elderly HF patients in the ICU.

Machine learning-based prediction of postoperative mortality in emergency colorectal surgery: A retrospective, multicenter cohort study using Tokushukai medical database

Background

Although prognostic factors associated with mortality in patients with emergency colorectal surgery have been identified, an accurate mortality risk assessment is still necessary to determine the range of therapeutic resources in accordance with the severity of patients. We established machine-learning models to predict in-hospital mortality for patients who had emergency colorectal surgery using clinical data at admission and attempted to identify prognostic factors associated with in-hospital mortality.

Methods

This retrospective cohort study included adult patients undergoing emergency colorectal surgery in 42 hospitals between 2012 and 2020. We employed logistic regression and three supervised machine-learning models: random forests, gradient-boosting decision trees (GBDT), and multilayer perceptron (MLP). The area under the receiver operating characteristics curve (AUROC) was calculated for each model. The Shapley additive explanations (SHAP) values are also calculated to identify the significant variables in GBDT.

Results

There were 8792 patients who underwent emergency colorectal surgery. As a result, the AUROC values of 0.742, 0.782, 0.814, and 0.768 were obtained for logistic regression, random forests, GBDT, and MLP. According to SHAP values, age, colorectal cancer, use of laparoscopy, and some laboratory variables, including serum lactate dehydrogenase serum albumin, and blood urea nitrogen, were significantly associated with in-hospital mortality.

Conclusion

We successfully generated a machine-learning prediction model, including GBDT, with the best prediction performance and exploited the potential for use in evaluating in-hospital mortality risk for patients who undergo emergency colorectal surgery.

Machine Learning for Mortality Prediction in Patients With Heart Failure With Mildly Reduced Ejection Fraction

Background Machine-learning-based prediction models (MLBPMs) have shown satisfactory performance in predicting clinical outcomes in patients with heart failure with reduced and preserved ejection fraction. However, their usefulness has yet to be fully elucidated in patients with heart failure with mildly reduced ejection fraction. This pilot study aims to evaluate the prediction performance of MLBPMs in a heart failure with mildly reduced ejection fraction cohort with long-term follow-up data. Methods and Results A total of 424 patients with heart failure with mildly reduced ejection fraction were enrolled in our study. The primary outcome was all-cause mortality. Two feature selection strategies were introduced for MLBPM development. The "All-in" (67 features) strategy was based on feature correlation, multicollinearity, and clinical significance. The other strategy was the CoxBoost algorithm with 10-fold cross-validation (17 features), which was based on the selection result of the "All-in" strategy. Six MLBPMs with 5-fold cross-validation based on the "All-in" and the CoxBoost algorithm with 10-fold cross-validation strategy were developed by the eXtreme Gradient Boosting, random forest, and support vector machine algorithms. The logistic regression model with 14 benchmark predictors was used as a reference model. During a median follow-up of 1008 (750, 1937) days, 121 patients met the primary outcome. Overall, MLBPMs outperformed the logistic model. The "All-in" eXtreme Gradient Boosting model had the best performance, with an accuracy of 85.4% and a precision of 70.3%. The area under the receiver-operating characteristic curve was 0.916 (95% CI, 0.887-0.945). The Brier score was 0.12. Conclusions The MLBPMs could significantly improve outcome prediction in patients with heart failure with mildly reduced ejection fraction, which would further optimize the management of these patients.

Acute heart failure with mildly reduced ejection fraction and myocardial infarction: a multi-institutional cohort study

BACKGROUND

Little research has been done on ischemic outcomes related to left ventricular ejection fraction (EF) in acute decompensated heart failure (ADHF).

METHODS

A retrospective cohort study was conducted between 2001 and 2021 using the Chang Gung Research Database. ADHF Patients discharged from hospitals between January 1, 2005, and December 31, 2019. Cardiovascular (CV) mortality and heart failure (HF) rehospitalization are the primary outcome components, along with all-cause mortality, acute myocardial infarction (AMI) and stroke.

RESULTS

A total of 12,852 ADHF patients were identified, of whom 2,222 (17.3%) had HFmrEF, the mean (SD) age was 68.5 (14.6) years, and 1,327 (59.7%) were males. In comparison with HFrEF and HFpEF patients, HFmrEF patients had a significant phenotype comorbid with diabetes, dyslipidemia, and ischemic heart disease. Patients with HFmrEF were more likely to experience renal failure, dialysis, and replacement. Both HFmrEF and HFrEF had similar rates of cardioversion and coronary interventions. There was an intermediate clinical outcome between HFpEF and HFrEF, but HFmrEF had the highest rate of AMI (HFpEF, 9.3%; HFmrEF, 13.6%; HFrEF, 9.9%). The AMI rates in HFmrEF were higher than those in HFpEF (AHR, 1.15; 95% Confidence Interval, 0.99 to 1.32) but not in HFrEF (AHR, 0.99; 95% Confidence Interval, 0.87 to 1.13).

CONCLUSION

Acute decompression in patients with HFmrEF increases the risk of myocardial infarction. The relationship between HFmrEF and ischemic cardiomyopathy, as well as optimal anti-ischemic treatment, requires further research on a large scale.

Analysis of Publication Activity and Research Trends in the Field of AI Medical Applications: Network Approach

Artificial intelligence (AI) has revolutionized numerous industries, including medicine. In recent years, the integration of AI into medical practices has shown great promise in enhancing the accuracy and efficiency of diagnosing diseases, predicting patient outcomes, and personalizing treatment plans. This paper aims at the exploration of the AI-based medicine research using network approach and analysis of existing trends based on PubMed. Our findings are based on the results of PubMed search queries and analysis of the number of papers obtained by the different search queries. Our goal is to explore how are the AI-based methods used in healthcare research, which approaches and techniques are the most popular, and to discuss the potential reasoning behind the obtained results. Using analysis of the co-occurrence network constructed using VOSviewer software, we detected the main clusters of interest in AI-based healthcare research. Then, we proceeded with the thorough analysis of publication activity in various categories of medical AI research, including research on different AI-based methods applied to different types of medical data. We analyzed the results of query processing in the PubMed database over the past 5 years obtained via a specifically designed strategy for generating search queries based on the thorough selection of keywords from different categories of interest. We provide a comprehensive analysis of existing applications of AI-based methods to medical data of different modalities, including the context of various medical fields and specific diseases that carry the greatest danger to the human population.

Data processing pipeline for cardiogenic shock prediction using machine learning

INTRODUCTION

Recent advances in machine learning provide new possibilities to process and analyse observational patient data to predict patient outcomes. In this paper, we introduce a data processing pipeline for cardiogenic shock (CS) prediction from the MIMIC III database of intensive cardiac care unit patients with acute coronary syndrome. The ability to identify high-risk patients could possibly allow taking pre-emptive measures and thus prevent the development of CS.

METHODS

We mainly focus on techniques for the imputation of missing data by generating a pipeline for imputation and comparing the performance of various multivariate imputation algorithms, including k-nearest neighbours, two singular value decomposition (SVD)-based methods, and Multiple Imputation by Chained Equations. After imputation, we select the final subjects and variables from the imputed dataset and showcase the performance of the gradient-boosted framework that uses a tree-based classifier for cardiogenic shock prediction.

RESULTS

We achieved good classification performance thanks to data cleaning and imputation (cross-validated mean area under the curve 0.805) without hyperparameter optimization.

CONCLUSION

We believe our pre-processing pipeline would prove helpful also for other classification and regression experiments.

U-shaped association between body mass index and ejection fraction in intensive care unit patients with heart failure

AIMS

There are limited data about the relationship between body mass index (BMI) and left ventricular ejection fraction (EF) in patients with heart failure (HF). The study aims to assess the correlation between BMI and left ventricular EF under HF conditions.

METHODS AND RESULTS

We derived the data from the Dryad Digital Repository for analysis, and the information of the original patients was obtained from the MIMIC-III database by the data uploader. We performed smooth curve and two piecewise linear regression analyses to evaluate the association between BMI and EF in HF patients. A total of 962 participants were included in this study, with age of 73.7 ± 13.5 years, and 475 participants were male (49.4%). The results of the smooth curve supported a U-shaped relationship between BMI and EF, and the inflection point was found to be a BMI of 23.3 kg/m2 in these HF patients. After adjusting for potential confounders, we found that EF decreased with increasing BMI up to the inflection point (β = -0.7, 95% CI -1.3 to -0.1, P = 0.028), whereas beyond the turning point, the relationship between EF and BMI showed a positive correlation (β = 0.2, 95% CI 0.1-0.3 P < 0.001). Importantly, ischaemic heart disease (interaction P = 0.0499) and hyperlipidaemia (interaction P = 0.0162) affected the association between BMI and EF in the lower BMI group (BMI < 23.3 kg/m2 ), although only diabetes mellitus (interaction P = 0.0255) altered the association between BMI and EF in the higher BMI group (BMI ≥ 23.3 kg/m2 ).

CONCLUSIONS

In addition to higher BMI, we also found that lower BMI is related to higher EF in intensive care unit patients with HF, supporting a U-shaped association between BMI and EF.

Mortality prediction among ICU inpatients based on MIMIC-III database results from the conditional medical generative adversarial network

Background and aims

Improved mortality prediction among intensive care unit (ICU) inpatients is a valuable and challenging task. Limited clinical data, especially with appropriate labels, are an important element restricting accurate predictions. Generative adversarial networks (GANs) are excellent generative models and have shown great potential for data simulation. However, there have been no relevant studies using GANs to predict mortality among ICU inpatients. In this study, we aim to evaluate the predictive performance of a variant of GAN called conditional medical GAN (c-med GAN) compared with some baseline models, including simplified acute physiology score II (SAPS II), support vector machine (SVM), and multilayer perceptron (MLP).

Methods

Data from a publicly available intensive care database, the Medical Information Mart for Intensive Care III (MIMIC-III) database (v1.4), were included in this study. The area under the precision-recall curve (PR-AUC), area under the receiver operating characteristic curve (ROC-AUC), and F1 score were used to evaluate the predictive performance. In addition, the size of the dataset was artificially reduced, and the performance of the c-med GAN was compared in different size datasets.

Results

The results showed that c-med GAN achieves the best PR-AUC, ROC-AUC, and F1 score compared with SAPS II, SVM, and MLP when training in the full MIMIC-III dataset. When the size of the dataset was reduced, the prediction performances of both MLP and c-med GAN were affected. However, the c-med GAN still outperformed MLP on smaller datasets and had less degradation.

Conclusion

The prediction of in-hospital mortality based on the c-med GAN for ICU patients showed better performance than the baseline models. Despite some inadequacies, this model may have a promising future in clinical applications which will be explored by further research.

Applying an Improved Stacking Ensemble Model to Predict the Mortality of ICU Patients with Heart Failure

Cardiovascular diseases have been identified as one of the top three causes of death worldwide, with onset and deaths mostly due to heart failure (HF). In ICU, where patients with HF are at increased risk of death and consume significant medical resources, early and accurate prediction of the time of death for patients at high risk of death would enable them to receive appropriate and timely medical care. The data for this study were obtained from the MIMIC-III database, where we collected vital signs and tests for 6699 HF patient during the first 24 h of their first ICU admission. In order to predict the mortality of HF patients in ICUs more precisely, an integrated stacking model is proposed and applied in this paper. In the first stage of dataset classification, the datasets were subjected to first-level classifiers using RF, SVC, KNN, LGBM, Bagging, and Adaboost. Then, the fusion of these six classifier decisions was used to construct and optimize the stacked set of second-level classifiers. The results indicate that our model obtained an accuracy of 95.25% and AUROC of 82.55% in predicting the mortality rate of HF patients, which demonstrates the outstanding capability and efficiency of our method. In addition, the results of this study also revealed that platelets, glucose, and blood urea nitrogen were the clinical features that had the greatest impact on model prediction. The results of this analysis not only improve the understanding of patients' conditions by healthcare professionals but allow for a more optimal use of healthcare resources.

Interpretable machine learning for 28-day all-cause in-hospital mortality prediction in critically ill patients with heart failure combined with hypertension: A retrospective cohort study based on medical information mart for intensive care database-IV and eICU databases

Background

Heart failure (HF) combined with hypertension is an extremely important cause of in-hospital mortality, especially for the intensive care unit (ICU) patients. However, under intense working pressure, the medical staff are easily overwhelmed by the large number of clinical signals generated in the ICU, which may lead to treatment delay, sub-optimal care, or even wrong clinical decisions. Individual risk stratification is an essential strategy for managing ICU patients with HF combined with hypertension. Artificial intelligence, especially machine learning (ML), can develop superior models to predict the prognosis of these patients. This study aimed to develop a machine learning method to predict the 28-day mortality for ICU patients with HF combined with hypertension.

Methods

We enrolled all critically ill patients with HF combined with hypertension in the Medical Information Mart for IntensiveCare Database-IV (MIMIC-IV, v.1.4) and the eICU Collaborative Research Database (eICU-CRD) from 2008 to 2019. Subsequently, MIMIC-IV was divided into training cohort and testing cohort in an 8:2 ratio, and eICU-CRD was designated as the external validation cohort. The least absolute shrinkage and selection operator (LASSO) Cox regression with internal tenfold cross-validation was used for data dimension reduction and identifying the most valuable predictive features for 28-day mortality. Based on its accuracy and area under the curve (AUC), the best model in the validation cohort was selected. In addition, we utilized the Shapley Additive Explanations (SHAP) method to highlight the importance of model features, analyze the impact of individual features on model output, and visualize an individual’s Shapley values.

Results

A total of 3,458 and 6582 patients with HF combined with hypertension in MIMIC-IV and eICU-CRD were included. The patients, including 1,756 males, had a median (Q1, Q3) age of 75 (65, 84) years. After selection, 22 out of a total of 58 clinical parameters were extracted to develop the machine-learning models. Among four constructed models, the Neural Networks (NN) model performed the best predictive performance with an AUC of 0.764 and 0.674 in the test cohort and external validation cohort, respectively. In addition, a simplified model including seven variables was built based on NN, which also had good predictive performance (AUC: 0.741). Feature importance analysis showed that age, mechanical ventilation (MECHVENT), chloride, bun, anion gap, paraplegia, rdw (RDW), hyperlipidemia, peripheral capillary oxygen saturation (SpO₂), respiratory rate, cerebrovascular disease, heart rate, white blood cell (WBC), international normalized ratio (INR), mean corpuscular hemoglobin concentration (MCHC), glucose, AIDS, mean corpuscular volume (MCV), N-terminal pro-brain natriuretic peptide (Npro. BNP), calcium, renal replacement therapy (RRT), and partial thromboplastin time (PTT) were the top 22 features of the NN model with the greatest impact. Finally, after hyperparameter optimization, SHAP plots were employed to make the NN-based model interpretable with an analytical description of how the constructed model visualizes the prediction of death.

Conclusion

We developed a predictive model to predict the 28-day mortality for ICU patients with HF combined with hypertension, which proved superior to the traditional logistic regression analysis. The SHAP method enables machine learning models to be more interpretable, thereby helping clinicians to better understand the reasoning behind the outcome and assess in-hospital outcomes for critically ill patients.

Explainable time-series deep learning models for the prediction of mortality, prolonged length of stay and 30-day readmission in intensive care patients

Background

In-hospital mortality, prolonged length of stay (LOS), and 30-day readmission are common outcomes in the intensive care unit (ICU). Traditional scoring systems and machine learning models for predicting these outcomes usually ignore the characteristics of ICU data, which are time-series forms. We aimed to use time-series deep learning models with the selective combination of three widely used scoring systems to predict these outcomes.

Materials and methods

A retrospective cohort study was conducted on 40,083 patients in ICU from the Medical Information Mart for Intensive Care-IV (MIMIC-IV) database. Three deep learning models, namely, recurrent neural network (RNN), gated recurrent unit (GRU), and long short-term memory (LSTM) with attention mechanisms, were trained for the prediction of in-hospital mortality, prolonged LOS, and 30-day readmission with variables collected during the initial 24 h after ICU admission or the last 24 h before discharge. The inclusion of variables was based on three widely used scoring systems, namely, APACHE II, SOFA, and SAPS II, and the predictors consisted of time-series vital signs, laboratory tests, medication, and procedures. The patients were randomly divided into a training set (80%) and a test set (20%), which were used for model development and model evaluation, respectively. The area under the receiver operating characteristic curve (AUC), sensitivity, specificity, and Brier scores were used to evaluate model performance. Variable significance was identified through attention mechanisms.

Results

A total of 33 variables for 40,083 patients were enrolled for mortality and prolonged LOS prediction and 36,180 for readmission prediction. The rates of occurrence of the three outcomes were 9.74%, 27.54%, and 11.79%, respectively. In each of the three outcomes, the performance of RNN, GRU, and LSTM did not differ greatly. Mortality prediction models, prolonged LOS prediction models, and readmission prediction models achieved AUCs of 0.870 ± 0.001, 0.765 ± 0.003, and 0.635 ± 0.018, respectively. The top significant variables co-selected by the three deep learning models were Glasgow Coma Scale (GCS), age, blood urea nitrogen, and norepinephrine for mortality; GCS, invasive ventilation, and blood urea nitrogen for prolonged LOS; and blood urea nitrogen, GCS, and ethnicity for readmission.

Conclusion

The prognostic prediction models established in our study achieved good performance in predicting common outcomes of patients in ICU, especially in mortality prediction. In addition, GCS and blood urea nitrogen were identified as the most important factors strongly associated with adverse ICU events.

Establishment of ICU Mortality Risk Prediction Models with Machine Learning Algorithm Using MIMIC-IV Database

Objective: The mortality rate of critically ill patients in ICUs is relatively high. In order to evaluate patients’ mortality risk, different scoring systems are used to help clinicians assess prognosis in ICUs, such as the Acute Physiology and Chronic Health Evaluation III (APACHE III) and the Logistic Organ Dysfunction Score (LODS). In this research, we aimed to establish and compare multiple machine learning models with physiology subscores of APACHE III—namely, the Acute Physiology Score III (APS III)—and LODS scoring systems in order to obtain better performance for ICU mortality prediction. Methods: A total number of 67,748 patients from the Medical Information Database for Intensive Care (MIMIC-IV) were enrolled, including 7055 deceased patients, and the same number of surviving patients were selected by the random downsampling technique, for a total of 14,110 patients included in the study. The enrolled patients were randomly divided into a training dataset (n = 9877) and a validation dataset (n = 4233). Fivefold cross-validation and grid search procedures were used to find and evaluate the best hyperparameters in different machine learning models. Taking the subscores of LODS and the physiology subscores that are part of the APACHE III scoring systems as input variables, four machine learning methods of XGBoost, logistic regression, support vector machine, and decision tree were used to establish ICU mortality prediction models, with AUCs as metrics. AUCs, specificity, sensitivity, positive predictive value, negative predictive value, and calibration curves were used to find the best model. Results: For the prediction of mortality risk in ICU patients, the AUC of the XGBoost model was 0.918 (95%CI, 0.915–0.922), and the AUCs of logistic regression, SVM, and decision tree were 0.872 (95%CI, 0.867–0.877), 0.872 (95%CI, 0.867–0.877), and 0.852 (95%CI, 0.847–0.857), respectively. The calibration curves of logistic regression and support vector machine performed better than the other two models in the ranges 0–40% and 70%–100%, respectively, while XGBoost performed better in the range of 40–70%. Conclusions: The mortality risk of ICU patients can be better predicted by the characteristics of the Acute Physiology Score III and the Logistic Organ Dysfunction Score with XGBoost in terms of ROC curve, sensitivity, and specificity. The XGBoost model could assist clinicians in judging in-hospital outcome of critically ill patients, especially in patients with a more uncertain survival outcome.

Predictive Value of Arterial Blood Lactic Acid Concentration on the Risk of in-Hospital All-Cause Death in Patients with Acute Heart Failure

The study aims to examine the predictive value of arterial blood lactic acid concentration for in-hospital all-cause mortality in the intensive care unit (ICU) for patients with acute heart failure (AHF). We retrospectively analyzed the clinical data of 7558 AHF patients in the Medical Information Mart for Intensive Care IV (MIMIC-IV) database. The exposure variable of the present study was arterial blood lactic acid concentration and the outcome variable was in-hospital all-cause death. The patients were divided into those who survived (n = 6792) and those who died (n = 766). The multivariate logistic regression model, restricted cubic spline (RCS) plot, and subgroup analysis were used to evaluate the association between lactic acid and in-hospital all-cause mortality. In addition, receiver operating curve (ROC) analysis also was performed. Finally, we further explore the association between NT-proBNP and lactic acid and in-hospital all-cause mortality. Compared with the lowest quartiles, the odds ratios with 95% confidence intervals for in-hospital all-cause mortality across the quartiles were 1.46 (1.07-2.00), 1.48 (1.09-2.00), and 2.36 (1.73-3.22) for lactic acid, and in-hospital all-cause mortality was gradually increased with lactic acid levels increasing (P for trend <0.05). The RCS plot revealed a positive and linear connection between lactic acid and in-hospital all-cause mortality. A combination of lactic acid concentration and the Simplified Acute Physiology Score (SAPS) II may improve the predictive value of in-hospital all-cause mortality in patients with AHF (AUC = 0.696). Among subgroups, respiratory failure interacted with an association between lactic acid and in-hospital all-cause mortality (P for interaction <0.05). The correlation heatmap revealed that NT-proBNP was positively correlated with lactic acid (r = 0.07) and positively correlated with in-hospital all-cause mortality (r = 0.18). There was an inverse L-shaped curve relationship between NT-proBNP and in-hospital all-cause mortality, respectively. Mediation analysis suggested that a positive relationship between lactic acid and in-hospital all-cause death was mediated by NT-proBNP. For AHF patients in the ICU, the arterial blood lactic acid concentration during hospitalization was a significant independent predictor of in-hospital all-cause mortality. The combination of lactic acid and SAPS II can improve the predictive value of the risk of in-hospital all-cause mortality in patients with AHF.

Prediction of in-hospital mortality of Clostriodiodes difficile infection using critical care database: a big data-driven, machine learning approach

RESEARCH OBJECTIVES

Clostriodiodes difficile infection (CDI) is a major cause of healthcare-associated diarrhoea with high mortality. There is a lack of validated predictors for severe outcomes in CDI. The aim of this study is to derive and validate a clinical prediction tool for CDI in-hospital mortality using a large critical care database.

METHODOLOGY

The demographics, clinical parameters, laboratory results and mortality of CDI were extracted from the Medical Information Mart for Intensive Care-III (MIMIC-III) database. We subsequently trained three machine learning models: logistic regression (LR), random forest (RF) and gradient boosting machine (GBM) to predict in-hospital mortality. The individual performances of the models were compared against current severity scores (Clostridiodes difficile Associated Risk of Death Score (CARDS) and ATLAS (Age, Treatment with systemic antibiotics, leukocyte count, Albumin and Serum creatinine as a measure of renal function) by calculating area under receiver operating curve (AUROC). We identified factors associated with higher mortality risk in each model.

SUMMARY OF RESULTS

From 61 532 intensive care unit stays in the MIMIC-III database, there were 1315 CDI cases. The mortality rate for CDI in the study cohort was 18.33%. AUROC was 0.69 (95% CI, 0.60 to 0.76) for LR, 0.71 (95% CI, 0.62 to 0.77) for RF and 0.72 (95% CI, 0.64 to 0.78) for GBM, while previously AUROC was 0.57 (95% CI, 0.51 to 0.65) for CARDS and 0.63 (95% CI, 0.54 to 0.70) for ATLAS. Albumin, lactate and bicarbonate were significant mortality factors for all the models. Free calcium, potassium, white blood cell, urea, platelet and mean blood pressure were present in at least two of the three models.

CONCLUSION

Our machine learning derived CDI in-hospital mortality prediction model identified pertinent factors that can assist critical care clinicians in identifying patients at high risk of dying from CDI.