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

Dexmedetomidine administration reduced mortality in patients with acute respiratory distress syndrome: a propensity score-matched cohort analysis

Background

Acute respiratory distress syndrome (ARDS) continues to pose significant difficulties due to the scarcity of successful preventative and therapeutic measures. Recent clinical trials and experimental research have confirmed the lung-protective and anti-inflammatory properties of dexmedetomidine. The objective of this study was to examine the relationship between the use of dexmedetomidine and mortality outcomes in ICU patients with ARDS.

Methods

This study retrospectively examined data from the Medical Information Mart for Intensive Care (MIMIC) IV, focusing on individuals diagnosed with ARDS. The primary endpoint was the occurrence of death within 28 days after entering the ICU. To ensure a balanced cohort, we applied propensity score matching at a 1:1 ratio. Additionally, multivariable analysis was performed to mitigate the effects of confounding factors.

Results

In this study, a cohort comprising 612 patients diagnosed with ARDS was investigated. Analysis using both univariate and multivariate Cox regression indicated significantly reduced 28-day and 90-day mortality rates in patients administered dexmedetomidine compared to those who were not given this treatment. Following adjustments for potential confounders using propensity score matching, these results were confirmed to be robust.

Conclusion

The results indicate an association between the administration of dexmedetomidine and lower mortality rates among severely ill ARDS patients. However, this result should be interpreted with cause because of a lot of missing data of potential risk factors for clinical outcomes. Nonetheless, it is imperative to perform further randomized controlled trials to corroborate this finding.

Inflammation-Driven Prognosis in Advanced Heart Failure: A Machine Learning-Based Risk Prediction Model for One-Year Mortality

Background

To develop a machine learning (ML)-based prediction model focused on the one-year mortality risk in patients with advanced heart failure (AdHF), aiming to improve prediction accuracy by integrating inflammatory biomarkers and clinical parameters, assist clinical decision-making, and enhance patient outcomes.

Methods

A retrospective cohort study. Data were obtained from the electronic medical records system of the Affiliated Hospital of Xuzhou Medical University. AdHF patients admitted to the ICU and cardiology department from January 2015 to December 2023 were included with a one-year follow-up. 52 variables potentially affecting prognosis were incorporated. The LASSO algorithm was used for feature selection and dimensionality reduction. Data were split into training and validation sets. Seven ML algorithms were applied to build and evaluate models. The SHAP method was used for model analysis and a dynamic nomogram was created.

Results

The study included 715 AdHF patients. The random forest (RF) model performed best, with an area under the curve (AUC) of 0.83 (95% confidence interval: 0.77-0.88), an accuracy of 0.72, a sensitivity of 0.74, and an F1 score of 0.73. Key predictors of one-year mortality risk included Beta blockers, ACEI/ARB/ARNI, BNP, CRP, NLR, AF, MI, NYHA class, and age. SHAP analysis revealed that elevated CRP, NLR, and age were associated with increased risk, while Beta blockers, ACEI/ARB/ARNI, and lower BNP values were associated with reduced risk. An online dynamic nomogram was developed to provide personalized risk predictions based on patient-specific conditions.

Conclusion

A successful ML-based prediction model was developed to accurately predict the one-year mortality risk in AdHF patients, with inflammation-driven factors being significant. The RF model integrating clinical features and inflammatory markers showed excellent performance and could assist clinical decision-making. Future research should conduct larger, multi-center, and prospective studies to further validate these findings.

Multimodal deep learning for predicting in-hospital mortality in heart failure patients using longitudinal chest X-rays and electronic health records

Amid an aging global population, heart failure has become a leading cause of hospitalization among older people. Its high prevalence and mortality rates underscore the importance of accurate mortality prediction for swift disease progression assessment and better patient outcomes. The evolution of artificial intelligence (AI) presents new avenues for predicting heart failure mortality. Yet current research has predominantly leveraged structured data and unstructured clinical notes from electronic health records (EHR), underutilizing the prognostic value of chest X-rays (CXRs). This study aims to harness deep learning methodologies to explore the feasibility of enhancing the precision of predicting in-hospital all-cause mortality in heart failure patients using CXRs data. We propose a novel multimodal deep learning network based on the spatially and temporally decoupled Transformer (MN-STDT) for in-hospital mortality prediction in heart failure by integrating longitudinal CXRs and structured EHR data. The MN-STDT captures spatial and temporal information from CXRs through a Hybrid Spatial Encoder and a Distance-Aware Temporal Encoder, ultimately fusing features from both modalities for predictive modeling. Initial pre-training of the spatial encoder was conducted on CheXpert, followed by full model training and evaluation on the MIMIC-IV and MIMIC-CXR datasets for mortality prediction tasks. In a comprehensive view, the MN-STDT demonstrated the best performance, with an AUC-ROC of 0.8620, surpassing all baseline models. Comparative analysis revealed that the AUC-ROC of the multimodal model (0.8620) was significantly higher than that of models using only structured data (0.8166) or chest X-ray data alone (0.7479). This study demonstrates the value of CXRs in the prognosis of heart failure, showing that the combination of longitudinal CXRs with structured EHR data can significantly improve the accuracy of mortality prediction in heart failure. Feature importance analysis based on SHAP provides interpretable decision support, paving the way for potential clinical applications.

Construction of risk prediction model and risk score table for infant heart failure hospital death based on white blood cell count to total protein ratio

The study analyzed 544 infants (1-36 months) with heart failure, aiming to correlate WBC/TP ratio with mortality and create a predictive model. Lasso regression identified significant mortality-associated clinical indices; logistic regression then built a death prediction model with an AUC of 0.755. Calibration and DCA curves indicated model accuracy and clinical utility. Risk stratification revealed higher mortality in the high-risk group, emphasizing WBC/TP as predictors and the model's value for early high-risk patient identification in infant heart failure.

Explainable machine learning and online calculators to predict heart failure mortality in intensive care units

AIMS

This study aims to develop explainable machine learning models and clinical tools for predicting mortality in patients in the intensive care unit (ICU) with heart failure (HF).

METHODS

Patients diagnosed with HF who experienced their first ICU stay lasting between 24 h and 28 days were selected from the Medical Information Mart for Intensive Care IV (MIMIC-IV) database. The primary outcome was all-cause mortality within 28 days. Data analysis was performed using Python and R, with feature selection conducted via least absolute shrinkage and selection operator (LASSO) regression. Fifteen models were evaluated, and the most effective model was rendered explainable through the Shapley additive explanations (SHAP) approach. A nomogram was developed based on logistic regression to facilitate interpretation. For external validation, the eICU database was utilized.

RESULTS

After selection, the study included 2343 records, with 1808 surviving and 535 deceased patients. The median age of the study population was 70.00, with ~3/5 males (60.31%). The median length of stay in the ICU was 6.00 days. The median age of the survival group was younger than the non-survival group (69.00 vs. 73.00), and non-survival patients spent longer time in the ICU. Seventy-five features were initially selected, including basic information, vital signs, laboratory tests, haemodynamics and oxygen status. LASSO regression determined the shrinkage parameter α = 0.020, and 44 features were chosen for model construction. The linear discriminant analysis (LDA) model showed the best performance, and the accuracy reached 0.8354 in the training cohort and 0.8563 in the testing cohort. It showed satisfying area under the curve (AUC), recall, precision, F1 score, Cohen's kappa score and Matthew's correlation coefficient. The concordance index (c-index) reached 0.7972 in the training cohort and 0.8125 in the testing cohort. In external validation, the LDA model achieved approximately 0.9 in accuracy, precision, recall and F1 score, with an AUC of 0.79. Univariable analysis was performed in the training cohort. Features that differed significantly between the survival and non-survival groups were subjected to multiple logistic regression. The nomogram built on multiple logistic regression included 14 features and demonstrated excellent performance. The AUC of the nomogram is 0.852 in the training cohort, 0.855 in the internal validation cohort and 0.770 in the external validation cohort. The calibration curve showed good consistency.

CONCLUSIONS

The study developed an LDA and a nomogram model for predicting mortality in HF patients in the ICU. The SHAP approach was employed to elucidate the LDA model, enhancing its utility for clinicians. These models were made accessible online for clinical application.

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.

Prognostic value of anion gap for patients with heart failure: a systematic review and meta-analysis

BACKGROUND

Heart failure (HF) is among the cardiovascular diseases with high morbidity and mortality worldwide. Due to the high burden of HF, finding easy-to-use prognostic factors has become important. Studies have investigated the correlation between anion gap (AG) and the HF prognosis. In this systematic review and meta-analysis, we aimed to evaluate the association between AG association with HF prognosis.

METHODS

PubMed, Embase, Scopus, and the Web of Science were systematically searched for studies evaluating AG in HF prognosis. Standardized mean difference (SMD) and pooled hazard ratio (HR) in addition to 95% confidence intervals (CIs) were calculated using random-effect meta-analyses to compare survivors vs. non-survivors.

RESULTS

Nine studies were included in this systematic review. In a random-effect meta-analysis comparing AG levels in those who died and survivors, non-survivors had significantly higher levels of AG (SMD 0.57, 95% CI 0.42 to 0.71, P < 0.0001, I2 = 46.4%). Meta-analysis of HRs for assessment of mortality revealed that high AG levels had significantly higher hazards of mortality, compared with low AG group (HR 1.64, 95% CI 1.35 to 1.99, P < 0.0001). Finally, a study investigated the association between intensive care unit (ICU) length of stay and AG in patients with HF which showed no significant association.

CONCLUSION

This study found that higher AG levels are associated with higher mortality in patients with HF which could be used in clinical settings and for patient management due to its ease of measurement and calculation. If confirmed in future studies, using this easy-to-measure index in clinical settings could provide useful information for clinicians in determining the risk of HF patients.

CLINICAL TRIAL NUMBER

Not applicable.

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.

Effects of acetaminophen use on mortality of patients with acute respiratory distress syndrome: secondary data mining based on the MIMIC-IV database

BACKGROUND

Acetaminophen is a commonly used analgesic after surgery, and its impact on prognosis in patients with acute respiratory distress syndrome (ARDS) has not been studied. This study explores the association between the use of acetaminophen and the risk of mortality in patients with ARDS.

METHODS

In this retrospective cohort study, 3,227 patients with ARDS who had or had not received acetaminophen were obtained from the Medical Information Mart for Intensive Care IV, patients were divided into acetaminophen and non- acetaminophen groups. In-hospital mortality of ARDS patients was considered as primary end point. We used univariate and multivariate Cox regression analyses to assess the relationship of acetaminophen use and in-hospital mortality in patients with ARDS. Subgroup analysis was performed according to age, gender, and severity of ARDS.

RESULTS

Of the total patients, 2,438 individuals were identified as acetaminophen users. The median duration of follow-up was 10.54 (5.57, 18.82) days. The results showed that the acetaminophen use was associated with a decreased risk of in-hospital mortality [hazard ratio (HR) = 0.67, 95% confidence interval (CI): 0.57-0.78]. Across various subgroups of patients with ARDS based on age, gender, and severity, acetaminophen use exhibited an association with reduced risk of in-hospital mortality.

CONCLUSION

Acetaminophen use was associated with in-hospital mortality of patients with ARDS. Acetaminophen therapy may represent a promising therapeutic option for ARDS patients and warrants further investigation.

Accounting for Red Cell Distribution Width Improves Risk Stratification by Commonly Used Mortality/Deterioration Risk Scores in Adult Patients Hospitalized Due to COVID-19

Higher red blood cell distribution width (RDW) levels have gained attention in the prognostication of many chronic metabolic and malignant diseases, as well as coronavirus disease 2019 (COVID-19). We aimed to evaluate whether accounting for RDW might contribute to risk stratification when added to commonly used risk scoring systems in adult COVID-19 patients. We retrospectively analyzed a cohort of 3212 non-critical COVID-19 patients hospitalized in a tertiary-level institution from March 2020 to June 2021. Admission RDW values were considered normal if they were ≤14.5% in males or ≤16.1% in females. The Modified Early Warning Score (MEWS), International Severe Acute Respiratory and Emerging Infections Consortium Coronavirus Clinical Characterisation Consortium score (ISARIC 4C), and Veterans Health Administration COVID-19 (VACO) index were evaluated as prognostic scores. RDW exceeded the upper limit in 628 (19.6%) of the patients. When RDW was accounted for, risks of the predicted outcomes were considerably different within the same MEWS, 4C score, and VACO index levels. The same patterns applied equally to patients who started, and those who did not start, remdesivir before deterioration. RDW may be a useful tool for stratifying risk when considered on top of commonly used prognostic scores in non-critical COVID-19 patients.

A review of evaluation approaches for explainable AI with applications in cardiology

Explainable artificial intelligence (XAI) elucidates the decision-making process of complex AI models and is important in building trust in model predictions. XAI explanations themselves require evaluation as to accuracy and reasonableness and in the context of use of the underlying AI model. This review details the evaluation of XAI in cardiac AI applications and has found that, of the studies examined, 37% evaluated XAI quality using literature results, 11% used clinicians as domain-experts, 11% used proxies or statistical analysis, with the remaining 43% not assessing the XAI used at all. We aim to inspire additional studies within healthcare, urging researchers not only to apply XAI methods but to systematically assess the resulting explanations, as a step towards developing trustworthy and safe models.

Supplementary Information

The online version contains supplementary material available at 10.1007/s10462-024-10852-w.

Development and Validation of Multimodal Models to Predict the 30-Day Mortality of ICU Patients Based on Clinical Parameters and Chest X-Rays

We aimed to develop and validate multimodal ICU patient prognosis models that combine clinical parameters data and chest X-ray (CXR) images. A total of 3798 subjects with clinical parameters and CXR images were extracted from the Medical Information Mart for Intensive Care IV (MIMIC-IV) database and an external hospital (the test set). The primary outcome was 30-day mortality after ICU admission. Automated machine learning (AutoML) and convolutional neural networks (CNNs) were used to construct single-modal models based on clinical parameters and CXR separately. An early fusion approach was used to integrate both modalities (clinical parameters and CXR) into a multimodal model named PrismICU. Compared to the single-modal models, i.e., the clinical parameter model (AUC = 0.80, F1-score = 0.43) and the CXR model (AUC = 0.76, F1-score = 0.45) and the scoring system APACHE II (AUC = 0.83, F1-score = 0.77), PrismICU (AUC = 0.95, F1 score = 0.95) showed improved performance in predicting the 30-day mortality in the validation set. In the test set, PrismICU (AUC = 0.82, F1-score = 0.61) was also better than the clinical parameters model (AUC = 0.72, F1-score = 0.50), CXR model (AUC = 0.71, F1-score = 0.36), and APACHE II (AUC = 0.62, F1-score = 0.50). PrismICU, which integrated clinical parameters data and CXR images, performed better than single-modal models and the existing scoring system. It supports the potential of multimodal models based on structured data and imaging in clinical management.

Effect of different sedatives on the prognosis of patients with mechanical ventilation: a retrospective cohort study based on MIMIC-IV database

Aim

To compare the effects of midazolam, propofol, and dexmedetomidine monotherapy and combination therapy on the prognosis of intensive care unit (ICU) patients receiving continuous mechanical ventilation (MV).

Methods

11,491 participants from the Medical Information Mart for Intensive Care (MIMIC)-IV database 2008-2019 was included in this retrospective cohort study. The primary outcome was defined as incidence of ventilator-associated pneumonia (VAP), in-hospital mortality, and duration of MV. Univariate and multivariate logistic regression analyses were utilized to evaluate the association between sedation and the incidence of VAP. Univariate and multivariate Cox analyses were performed to investigate the correlation between sedative therapy and in-hospital mortality. Additionally, univariate and multivariate linear analyses were conducted to explore the relationship between sedation and duration of MV.

Results

Compared to patients not receiving these medications, propofol alone, dexmedetomidine alone, combination of midazolam and dexmedetomidine, combination of propofol and dexmedetomidine, combination of midazolam, propofol and dexmedetomidine were all association with an increased risk of VAP; dexmedetomidine alone, combination of midazolam and dexmedetomidine, combination of propofol and dexmedetomidine, combination of midazolam, propofol and dexmedetomidine may be protective factor for in-hospital mortality, while propofol alone was risk factor. There was a positive correlation between all types of tranquilizers and the duration of MV. Taking dexmedetomidine alone as the reference, all other drug groups were found to be associated with an increased risk of in-hospital mortality. The administration of propofol alone, in combination with midazolam and dexmedetomidine, in combination with propofol and dexmedetomidine, in combination with midazolam, propofol and dexmedetomidine were associated with an increased risk of VAP compared to the use of dexmedetomidine alone.

Conclusion

Dexmedetomidine alone may present as a favorable prognostic option for ICU patients with mechanical ventilation MV.

Interpretability-based machine learning for predicting the risk of death from pulmonary inflammation in Chinese intensive care unit patients

Objective

The objective of this research was to create a machine learning predictive model that could be easily interpreted in order to precisely determine the risk of premature death in patients receiving intensive care after pulmonary inflammation.

Methods

In this study, information from the China intensive care units (ICU) Open Source database was used to examine data from 2790 patients who had infections between January 2019 and December 2020. A 7:3 ratio was used to randomly assign the whole patient population to training and validation groups. This study used six machine learning techniques: logistic regression, random forest, gradient boosting tree, extreme gradient boosting tree (XGBoost), multilayer perceptron, and K-nearest neighbor. A cross-validation grid search method was used to search the parameters in each model. Eight metrics were used to assess the models' performance: accuracy, precision, recall, F1 score, area under the curve (AUC) value, Brier score, Jordon's index, and calibration slope. The machine methods were ranked based on how well they performed in each of these metrics. The best-performing models were selected for interpretation using both the Shapley Additive exPlanations (SHAP) and Local interpretable model-agnostic explanations (LIME) interpretable techniques.

Results

A subset of the study cohort's patients (120/1668, or 7.19%) died in the hospital following screening for inclusion and exclusion criteria. Using a cross-validated grid search to evaluate the six machine learning techniques, XGBoost showed good discriminative ability, achieving an accuracy score of 0.889 (0.874-0.904), precision score of 0.871 (0.849-0.893), recall score of 0.913 (0.890-0.936), F1 score of 0.891 (0.876-0.906), and AUC of 0.956 (0.939-0.973). Additionally, XGBoost exhibited excellent performance with a Brier score of 0.050, Jordon index of 0.947, and calibration slope of 1.074. It was also possible to create an interactive internet page using the XGBoost model.

Conclusion

By identifying patients at higher risk of early mortality, machine learning-based mortality risk prediction models have the potential to significantly improve patient care by directing clinical decision making and enabling early detection of survival and mortality issues in patients with pulmonary inflammation disease.

Explainable machine learning for predicting neurological outcome in hemorrhagic and ischemic stroke patients in critical care

Aim

The objective of this study is to develop accurate machine learning (ML) models for predicting the neurological status at hospital discharge of critically ill patients with hemorrhagic and ischemic stroke and identify the risk factors associated with the neurological outcome of stroke, thereby providing healthcare professionals with enhanced clinical decision-making guidance.

Materials and methods

Data of stroke patients were extracted from the eICU Collaborative Research Database (eICU-CRD) for training and testing sets and the Medical Information Mart for Intensive Care IV (MIMIC IV) database for external validation. Four machine learning models, namely gradient boosting classifier (GBC), logistic regression (LR), multi-layer perceptron (MLP), and random forest (RF), were used for prediction of neurological outcome. Furthermore, shapley additive explanations (SHAP) algorithm was applied to explain models visually.

Results

A total of 1,216 hemorrhagic stroke patients and 954 ischemic stroke patients from eICU-CRD and 921 hemorrhagic stroke patients 902 ischemic stroke patients from MIMIC IV were included in this study. In the hemorrhagic stroke cohort, the LR model achieved the highest area under curve (AUC) of 0.887 in the test cohort, while in the ischemic stroke cohort, the RF model demonstrated the best performance with an AUC of 0.867 in the test cohort. Further analysis of risk factors was conducted using SHAP analysis and the results of this study were converted into an online prediction tool.

Conclusion

ML models are reliable tools for predicting hemorrhagic and ischemic stroke neurological outcome and have the potential to improve critical care of stroke patients. The summarized risk factors obtained from SHAP enable a more nuanced understanding of the reasoning behind prediction outcomes and the optimization of the treatment strategy.

An interpretable machine learning model for predicting 28-day mortality in patients with sepsis-associated liver injury

Sepsis-Associated Liver Injury (SALI) is an independent risk factor for death from sepsis. The aim of this study was to develop an interpretable machine learning model for early prediction of 28-day mortality in patients with SALI. Data from the Medical Information Mart for Intensive Care (MIMIC-IV, v2.2, MIMIC-III, v1.4) were used in this study. The study cohort from MIMIC-IV was randomized to the training set (0.7) and the internal validation set (0.3), with MIMIC-III (2001 to 2008) as external validation. The features with more than 20% missing values were deleted and the remaining features were multiple interpolated. Lasso-CV that lasso linear model with iterative fitting along a regularization path in which the best model is selected by cross-validation was used to select important features for model development. Eight machine learning models including Random Forest (RF), Logistic Regression, Decision Tree, Extreme Gradient Boost (XGBoost), K Nearest Neighbor, Support Vector Machine, Generalized Linear Models in which the best model is selected by cross-validation (CV_glmnet), and Linear Discriminant Analysis (LDA) were developed. Shapley additive interpretation (SHAP) was used to improve the interpretability of the optimal model. At last, a total of 1043 patients were included, of whom 710 were from MIMIC-IV and 333 from MIMIC-III. Twenty-four clinically relevant parameters were selected for model construction. For the prediction of 28-day mortality of SALI in the internal validation set, the area under the curve (AUC (95% CI)) of RF was 0.79 (95% CI: 0.73-0.86), and which performed the best. Compared with the traditional disease severity scores including Oxford Acute Severity of Illness Score (OASIS), Sequential Organ Failure Assessment (SOFA), Simplified Acute Physiology Score II (SAPS II), Logistic Organ Dysfunction Score (LODS), Systemic Inflammatory Response Syndrome (SIRS), and Acute Physiology Score III (APS III), RF also had the best performance. SHAP analysis found that Urine output, Charlson Comorbidity Index (CCI), minimal Glasgow Coma Scale (GCS_min), blood urea nitrogen (BUN) and admission_age were the five most important features affecting RF model. Therefore, RF has good predictive ability for 28-day mortality prediction in SALI. Urine output, CCI, GCS_min, BUN and age at admission(admission_age) within 24 h after intensive care unit(ICU) admission contribute significantly to model prediction.

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

ABSTRACT

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

Leveraging artificial intelligence to advance implementation science: potential opportunities and cautions

BACKGROUND

The field of implementation science was developed to address the significant time delay between establishing an evidence-based practice and its widespread use. Although implementation science has contributed much toward bridging this gap, the evidence-to-practice chasm remains a challenge. There are some key aspects of implementation science in which advances are needed, including speed and assessing causality and mechanisms. The increasing availability of artificial intelligence applications offers opportunities to help address specific issues faced by the field of implementation science and expand its methods.

MAIN TEXT

This paper discusses the many ways artificial intelligence can address key challenges in applying implementation science methods while also considering potential pitfalls to the use of artificial intelligence. We answer the questions of "why" the field of implementation science should consider artificial intelligence, for "what" (the purpose and methods), and the "what" (consequences and challenges). We describe specific ways artificial intelligence can address implementation science challenges related to (1) speed, (2) sustainability, (3) equity, (4) generalizability, (5) assessing context and context-outcome relationships, and (6) assessing causality and mechanisms. Examples are provided from global health systems, public health, and precision health that illustrate both potential advantages and hazards of integrating artificial intelligence applications into implementation science methods. We conclude by providing recommendations and resources for implementation researchers and practitioners to leverage artificial intelligence in their work responsibly.

CONCLUSIONS

Artificial intelligence holds promise to advance implementation science methods ("why") and accelerate its goals of closing the evidence-to-practice gap ("purpose"). However, evaluation of artificial intelligence's potential unintended consequences must be considered and proactively monitored. Given the technical nature of artificial intelligence applications as well as their potential impact on the field, transdisciplinary collaboration is needed and may suggest the need for a subset of implementation scientists cross-trained in both fields to ensure artificial intelligence is used optimally and ethically.

Predicting the risk of mortality and rehospitalization in heart failure patients: A retrospective cohort study by machine learning approach

BACKGROUND

Heart failure (HF) is a global problem, affecting more than 26 million people worldwide. This study evaluated the performance of 10 machine learning (ML) algorithms and chose the best algorithm to predict mortality and readmission of HF patients by using The Fasa Registry on Systolic HF (FaRSH) database.

HYPOTHESIS

ML algorithms may better identify patients at increased risk of HF readmission or death with demographic and clinical data.

METHODS

Through comprehensive evaluation, the best-performing model was used for prediction. Finally, all the trained models were applied to the test data, which included 20% of the total data. For the final evaluation and comparison of the models, five metrics were used: accuracy, F1-score, sensitivity, specificity and Area Under Curve (AUC).

RESULTS

Ten ML algorithms were evaluated. The CatBoost (CAT) algorithm uses a series of decision tree models to create a nonlinear model, and this CAT algorithm performed the best of the 10 models studied. According to the three final outcomes from this study, which involved 2488 participants, 366 (14.7%) of the patients were readmitted to the hospital, 97 (3.9%) of the patients died within 1 month of the follow-up, and 342 (13.7%) of the patients died within 1 year of the follow-up. The most significant variables to predict the events were length of stay in the hospital, hemoglobin level, and family history of MI.

CONCLUSIONS

The ML-based risk stratification tool was able to assess the risk of 5-year all-cause mortality and readmission in patients with HF. ML could provide an explicit explanation of individualized risk prediction and give physicians an intuitive understanding of the influence of critical features in the model.

Development and validation of a novel combinatorial nomogram model to predict in-hospital deaths in heart failure patients

BACKGROUND

The purpose of this study was to develop a Nomogram model to identify the risk of all-cause mortality during hospitalization in patients with heart failure (HF).

METHODS

HF patients who had been registered in the Medical Information Mart for Intensive Care (MIMIC) III and IV databases were included. The primary outcome was the occurrence of all-cause mortality during hospitalization. Two Logistic Regression models (LR1 and LR2) were developed to predict in-hospital death for HF patients from the MIMIC-IV database. The MIMIC-III database were used for model validation. The area under the receiver operating characteristic curve (AUC) was used to compare the discrimination of each model. Calibration curve was used to assess the fit of each developed models. Decision curve analysis (DCA) was used to estimate the net benefit of the predictive model.

RESULTS

A total of 16,908 HF patients were finally enrolled through screening, of whom 2,283 (13.5%) presented with in-hospital death. Totally, 48 variables were included and analyzed in the univariate and multifactorial regression analysis. The AUCs for the LR1 and LR2 models in the test cohort were 0.751 (95% CI: 0.735∼0.767) and 0.766 (95% CI: 0.751-0.781), respectively. Both LR models performed well in the calibration curve and DCA process. Nomogram and online risk assessment system were used as visualization of predictive models.

CONCLUSION

A new risk prediction tool and an online risk assessment system were developed to predict mortality in HF patients, which performed well and might be used to guide clinical practice.

Acute kidney injury should not be neglected - optimization of quick Pitt bacteremia score for predicting mortality in critically ill patients with bloodstream infection: a retrospective cohort study

Background

Considering the therapeutic difficulties and mortality associated with bloodstream infection (BSI), it is essential to investigate other potential factors affecting mortality in critically ill patients with BSI and examine the utility of the quick Pitt bacteremia (qPitt) score to improve the survival rate.

Objectives

To improve the predictive accuracy of the qPitt scoring system by evaluating the five current components of qPitt and including other potential factors influencing mortality in critically ill patients with BSI.

Design

This was a retrospective cohort study.

Methods

Medical information from the Medical Information Mart for Intensive Care IV database was used in this retrospective cohort study. The risk factors associated with mortality were examined using a multivariate logistic regression model. The area under the receiver operating characteristic curve (AUC) was used to assess the discriminatory capability of the prediction models.

Results

In total, 1240 eligible critically ill patients with BSI were included. After adjustment for age, community-onset BSI, indwelling invasive lines, and Glasgow Coma Scale (GCS) ⩽ 8, acute kidney injury (AKI) was identified as a notable risk factor for 14-day mortality. Except for altered mental status, the four other main components of the original qPitt were significantly associated with 14-day mortality. Hence, we established a modified qPitt (m-qPitt) by adding AKI and replacing altered mental status with GCS ⩽ 8. The AUCs for m-qPitt and qPitt were 0.723 [95% confidence interval (CI): 0.683-0.759] and 0.708 (95% CI: 0.669-0.745) in predicting 14-day mortality, respectively. Moreover, m-qPitt also had acceptable performance and discrimination power [0.700 (95% CI: 0.666-0.732)] in predicting 28-day mortality.

Conclusion

AKI significantly influenced the survival of critically ill patients with BSIs. Compared with the original qPitt, our new m-qPitt was proven to have a better predictive performance for mortality in critically ill patients with BSI. Further studies should be conducted to validate the practicality of m-qPitt.

Shapely additive values can effectively visualize pertinent covariates in machine learning when predicting hypertension

Machine learning methods are widely used within the medical field to enhance prediction. However, little is known about the reliability and efficacy of these models to predict long-term medical outcomes such as blood pressure using lifestyle factors, such as diet. The authors assessed whether machine-learning techniques could accurately predict hypertension risk using nutritional information. A cross-sectional study using data from the National Health and Nutrition Examination Survey (NHANES) between January 2017 and March 2020. XGBoost was used as the machine-learning model of choice in this study due to its increased performance relative to other common methods within medical studies. Model prediction metrics (e.g., AUROC, Balanced Accuracy) were used to measure overall model efficacy, covariate Gain statistics (percentage each covariate contributes to the overall prediction) and SHapely Additive exPlanations (SHAP, method to visualize each covariate) were used to provide explanations to machine-learning output and increase the transparency of this otherwise cryptic method. Of a total of 9650 eligible patients, the mean age was 41.02 (SD = 22.16), 4792 (50%) males, 4858 (50%) female, 3407 (35%) White patients, 2567 (27%) Black patients, 2108 (22%) Hispanic patients, and 981 (10%) Asian patients. From evaluation of model gain statistics, age was found to be the single strongest predictor of hypertension, with a gain of 53.1%. Additionally, demographic factors such as poverty and Black race were also strong predictors of hypertension, with gain of 4.33% and 4.18%, respectively. Nutritional Covariates contributed 37% to the overall prediction: Sodium, Caffeine, Potassium, and Alcohol intake being significantly represented within the model. Machine Learning can be used to predict hypertension.

Association of inflammatory indicators with intensive care unit mortality in critically ill patients with coronary heart disease

Objective

Coronary heart disease (CHD) is one of the major cardiovascular diseases, a common chronic disease in the elderly and a major cause of disability and death in the world. Currently, intensive care unit (ICU) patients have a high probability of concomitant coronary artery disease, and the mortality of this category of patients in the ICU is receiving increasing attention. Therefore, the aim of this study was to verify whether the composite inflammatory indicators are significantly associated with ICU mortality in ICU patients with CHD and to develop a simple personalized prediction model.

Method

7115 patients from the Multi-Parameter Intelligent Monitoring in Intensive Care Database IV were randomly assigned to the training cohort (n = 5692) and internal validation cohort (n = 1423), and 701 patients from the eICU Collaborative Research Database served as the external validation cohort. The association between various inflammatory indicators and ICU mortality was determined by multivariate Logistic regression analysis and Cox proportional hazards model. Subsequently, a novel predictive model for mortality in ICU patients with CHD was developed in the training cohort and performance was evaluated in the internal and external validation cohorts.

Results

Various inflammatory indicators were demonstrated to be significantly associated with ICU mortality, 30-day ICU mortality, and 90-day ICU mortality in ICU patients with CHD by Logistic regression analysis and Cox proportional hazards model. The area under the curve of the novel predictive model for ICU mortality in ICU patients with CHD was 0.885 for the internal validation cohort and 0.726 for the external validation cohort. The calibration curve showed that the predicted probabilities of the model matched the actual observed probabilities. Furthermore, the decision curve analysis showed that the novel prediction model had a high net clinical benefit.

Conclusion

In ICU patients with CHD, various inflammatory indicators were independent risk factors for ICU mortality. We constructed a novel predictive model of ICU mortality risk in ICU patients with CHD that had great potential to guide clinical decision-making.

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.

Association of red blood cell distribution width with hospital admission and in-hospital mortality across all-cause adult emergency department visits

Objectives

To test the association between the initial red blood cell distribution width (RDW) value in the emergency department (ED) and hospital admission and, among those admitted, in-hospital mortality.

Materials and Methods

We perform a retrospective analysis of 210 930 adult ED visits with complete blood count results from March 2013 to February 2022. Primary outcomes were hospital admission and in-hospital mortality. Variables for each visit included demographics, comorbidities, vital signs, basic metabolic panel, complete blood count, and final diagnosis. The association of each outcome with the initial RDW value was calculated across 3 age groups (<45, 45-65, and >65) as well as across 374 diagnosis categories. Logistic regression (LR) and XGBoost models using all variables excluding final diagnoses were built to test whether RDW was a highly weighted and informative predictor for each outcome. Finally, simplified models using only age, sex, and vital signs were built to test whether RDW had additive predictive value.

Results

Compared to that of discharged visits (mean [SD]: 13.8 [2.03]), RDW was significantly elevated in visits that resulted in admission (15.1 [2.72]) and, among admissions, those resulting in intensive care unit stay (15.3 [2.88]) and/or death (16.8 [3.25]). This relationship held across age groups as well as across various diagnosis categories. An RDW >16 achieved 90% specificity for hospital admission, while an RDW >18.5 achieved 90% specificity for in-hospital mortality. LR achieved a test area under the curve (AUC) of 0.77 (95% confidence interval [CI] 0.77-0.78) for hospital admission and 0.85 (95% CI 0.81-0.88) for in-hospital mortality, while XGBoost achieved a test AUC of 0.90 (95% CI 0.89-0.90) for hospital admission and 0.96 (95% CI 0.94-0.97) for in-hospital mortality. RDW had high scaled weights and information gain for both outcomes and had additive value in simplified models predicting hospital admission.

Discussion

Elevated RDW, previously associated with mortality in myocardial infarction, pulmonary embolism, heart failure, sepsis, and COVID-19, is associated with hospital admission and in-hospital mortality across all-cause adult ED visits. Used alone, elevated RDW may be a specific, but not sensitive, test for both outcomes, with multivariate LR and XGBoost models showing significantly improved test characteristics.

Conclusions

RDW, a component of the complete blood count panel routinely ordered as the initial workup for the undifferentiated patient, may be a generalizable biomarker for acuity in the ED.

Association between admission hyperglycemia and postoperative pneumonia in geriatric patients with hip fractures

BACKGROUND

Admission hyperglycemia is a common phenomenon in the early stages of injury. This study aimed to determine the relationship between admission hyperglycemia and postoperative pneumonia in geriatric patients with hip fractures.

METHODS

A total of 600 geriatric patients admitted to Dandong Central Hospital with hip fractures were included. Patients were divided into four groups based on quartiles of admission blood glucose levels: Q1- Q4. Multivariable logistic regression and propensity score-matched analyses were conducted to calculate adjusted odds ratios (ORs) and 95% confidence intervals (CIs) for postoperative pneumonia. Receiver operating characteristic (ROC) curves were used to determine the cut-off value of admission hyperglycemia for predicting postoperative pneumonia.

RESULTS

The incidence of postoperative pneumonia was significantly higher among hyperglycemic patients than those with normal glucose levels (OR = 2.090, 95% CI: 1.135-3.846, p = 0.016). Admission hyperglycemia showed moderate predictive power, with an area under the ROC curve of 0.803. Furthermore, propensity score-matched analyses demonstrated that patients in the Q3 (OR = 4.250, 95% CI: 1.361-13.272, p = 0.013) and Q4 (OR = 4.667, 95% CI: 1.251-17.405, p = 0.022) quartiles had a significantly higher risk of postoperative pneumonia compared to patients in the Q1 quartile.

CONCLUSIONS

Admission hyperglycemia in elderly hip fracture patients increases the risk of postoperative pneumonia. This biomarker can aid clinical assessment and perioperative management.

The predictive values of admission characteristics for 28-day all-cause mortality in septic patients with diabetes mellitus: a study from the MIMIC database

Background

Septic patients with diabetes mellitus (DM) are more venerable to subsequent complications and the resultant increase in associated mortality. Therefore, it is important to make tailored clinical decisions for this subpopulation at admission.

Method

Data from large-scale real-world databases named the Medical Information Mart for Intensive Care Database (MIMIC) were reviewed. The least absolute selection and shrinkage operator (LASSO) was performed with 10 times cross-validation methods to select the optimal prognostic factors. Multivariate COX regression analysis was conducted to identify the independent prognostic factors and nomogram construction. The nomogram was internally validated via the bootstrapping method and externally validated by the MIMIC III database with receiver operating characteristic (ROC), calibration curves, decision curve analysis (DCA), and Kaplan-Meier curves for robustness check.

Results

A total of 3,291 septic patients with DM were included in this study, 2,227 in the MIMIC IV database and 1,064 in the MIMIC III database, respectively. In the training cohort, the 28-day all-cause mortality rate is 23.9% septic patients with DM. The multivariate Cox regression analysis reveals age (hazard ratio (HR)=1.023, 95%CI: 1.016-1.031, p<0.001), respiratory failure (HR=1.872, 95%CI: 1.554-2.254, p<0.001), Sequential Organ Failure Assessment score (HR=1.056, 95%CI: 1.018-1.094, p=0.004); base excess (HR=0.980, 95%CI: 0.967-0.992, p=0.002), anion gap (HR=1.100, 95%CI: 1.080-1.120, p<0.001), albumin (HR=0.679, 95%CI: 0.574-0.802, p<0.001), international normalized ratio (HR=1.087, 95%CI: 1.027-1.150, p=0.004), red cell distribution width (HR=1.056, 95%CI: 1.021-1.092, p=0.001), temperature (HR=0.857, 95%CI: 0.789-0.932, p<0.001), and glycosylated hemoglobin (HR=1.358, 95%CI: 1.320-1.401, p<0.001) at admission are independent prognostic factors for 28-day all-cause mortality of septic patients with DM. The established nomogram shows satisfied accuracy and clinical utility with AUCs of 0.870 in the internal validation and 0.830 in the external validation cohort as well as 0.820 in the septic shock subpopulation, which is superior to the predictive value of the single SOFA score.

Conclusion

Our results suggest that admission characteristics show an optimal prediction value for short-term mortality in septic patients with DM. The established model can support intensive care unit physicians in making better initial clinical decisions for this subpopulation.

Development and Validation of a Dynamic Nomogram for Predicting in-Hospital Mortality in Patients with Acute Pancreatitis: A Retrospective Cohort Study in the Intensive Care Unit

PURPOSE

The aim of this study is to develop and validate a predictive model for the prediction of in-hospital mortality in patients with acute pancreatitis (AP) based on the intensive care database.

PATIENTS AND METHODS

We analyzed the data of patients with AP in the Medical Information Mart for Intensive Care-IV (MIMIC-IV) database and Electronic Intensive Care Unit Collaborative Research Database (eICU-CRD). Then, patients from MIMIC-IV were divided into a development group and a validation group according to the ratio of 8:2, and eICU-CRD was assigned as an external validation group. Univariate logistic regression and least absolute shrinkage and selection operator regression were used for screening the best predictors, and multivariate logistic regression was used to establish a dynamic nomogram. We evaluated the discrimination, calibration, and clinical efficacy of the nomogram, and compared the performance of the nomogram with Acute Physiology and Chronic Health Evaluation II (APACHE-II) score and Bedside Index of Severity in AP (BISAP) score.

RESULTS

A total of 1030 and 514 patients with AP in MIMIC-IV database and eICU-CRD were included in the study. After stepwise analysis, 8 out of a total of 37 variables were selected to construct the nomogram. The dynamic nomogram can be obtained by visiting https://model.sci-inn.com/KangZou/. The area under receiver operating characteristic curve (AUC) of the nomogram was 0.859, 0.871, and 0.847 in the development, internal, and external validation set respectively. The nomogram had a similar performance with APACHE-II (AUC = 0.841, p = 0.537) but performed better than BISAP (AUC = 0.690, p = 0.001) score in the validation group. Moreover, the calibration curve presented a satisfactory predictive accuracy, and the decision curve analysis suggested great clinical application value of the nomogram.

CONCLUSION

Based on the results of internal and external validation, the nomogram showed favorable discrimination, calibration, and clinical practicability in predicting the in-hospital mortality of patients with AP.

Association of thiamine administration and prognosis in critically ill patients with heart failure

Background: Thiamine deficiency is common in patients with heart failure, and thiamine supplement can benefit these patients. However, the association between thiamine administration and prognosis among critically ill patients with heart failure remains unclear. Thus, this study aims to prove the survival benefit of thiamine use in critically ill patients with heart failure. Methods: A retrospective cohort analysis was performed on the basis of the Medical Information Mart of Intensive Care-Ⅳ database. Critically ill patients with heart failure were divided into the thiamine and non-thiamine groups depending on whether they had received thiamine therapy or not during hospitalization. The association between thiamine supplement and in-hospital mortality was assessed by using the Kaplan-Meier (KM) method and Cox proportional hazard models. A 1:1 nearest propensity-score matching (PSM) and propensity score-based inverse probability of treatment weighting (IPW) were also performed to ensure the robustness of the findings. Results: A total of 7,021 patients were included in this study, with 685 and 6,336 in the thiamine and non-thiamine groups, respectively. The kaplan-meier survival curves indicated that the thiamine group had a lower in-hospital mortality than the none-thiamine group. After adjusting for various confounders, the Cox regression models showed significant beneficial effects of thiamine administration on in-hospital mortality among critically ill patients with heart failure with a hazard ratio of 0.78 (95% confidence interval: 0.67-0.89) in the fully adjusted model. propensity-score matching and probability of treatment weighting analyses also achieved consistent results. Conclusion: Thiamine supplement is associated with a decreased risk of in-hospital mortality in critically ill patients with heart failure who are admitted to the ICU. Further multicenter and well-designed randomized controlled trials with large sample sizes are necessary to validate this finding.