Prediction of short-term mortality in acute heart failure patients using minimal electronic health record data

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.

Predicting the risk of mortality in ICU patients based on dynamic graph attention network of patient similarity

Predicting the risk of mortality of hospitalized patients in the ICU is essential for timely identification of high-risk patients and formulate and adjustment of treatment strategies when patients are hospitalized. Traditional machine learning methods usually ignore the similarity between patients and make it difficult to uncover the hidden relationships between patients, resulting in poor accuracy of prediction models. In this paper, we propose a new model named PS-DGAT to solve the above problem. First, we construct a patient-weighted similarity network by calculating the similarity of patient clinical data to represent the similarity relationship between patients; second, we fill in the missing features and reconstruct the patient similarity network based on the data of neighboring patients in the network; finally, from the reconstructed patient similarity network after feature completion, we use the dynamic attention mechanism to extract and learn the structural features of the nodes to obtain a vector representation of each patient node in the low-dimensional embedding The vector representation of each patient node in the low-dimensional embedding space is used to achieve patient mortality risk prediction. The experimental results show that the accuracy is improved by about 1.8% compared with the basic GAT and about 8% compared with the traditional machine learning methods.

Data Modeling Using Vital Sign Dynamics for In-hospital Mortality Classification in Patients with Acute Coronary Syndrome

OBJECTIVES

This study compared feature selection by machine learning or expert recommendation in the performance of classification models for in-hospital mortality among patients with acute coronary syndrome (ACS) who underwent percutaneous coronary intervention (PCI).

METHODS

A dataset of 1,123 patients with ACS who underwent PCI was analyzed. After assigning 80% of instances to the training set through random splitting, we performed feature scaling and resampling with the synthetic minority over-sampling technique and Tomek link method. We compared two feature selection.

METHODS

recursive feature elimination with cross-validation (RFECV) and selection by interventional cardiologists. We used five simple models: support vector machine (SVM), random forest, decision tree, logistic regression, and artificial neural network. The performance metrics were accuracy, recall, and the false-negative rate, measured with 10-fold cross-validation in the training set and validated in the test set.

RESULTS

Patients' mean age was 66.22 ± 12.88 years, and 33.63% had ST-elevation ACS. Fifteen of 34 features were selected as important with the RFECV method, while the experts chose 11 features. All models with feature selection by RFECV had higher accuracy than the models with expert-chosen features. In the training set, the random forest model had the highest accuracy (0.96 ± 0.01) and recall (0.97 ± 0.02). After validation in the test set, the SVM model displayed the highest accuracy (0.81) and a recall of 0.61.

CONCLUSIONS

Models with feature selection by RFECV had higher accuracy than those with feature selection by experts in identifying patients with ACS at high risk for in-hospital mortality.

Predicting heart failure in-hospital mortality by integrating longitudinal and category data in electronic health records

Heart failure is a life-threatening syndrome that is diagnosed in 3.6 million people worldwide each year. We propose a deep fusion learning model (DFL-IMP) that uses time series and category data from electronic health records to predict in-hospital mortality in patients with heart failure. We considered 41 time series features (platelets, white blood cells, urea nitrogen, etc.) and 17 category features (gender, insurance, marital status, etc.) as predictors, all of which were available within the time of the patient's last hospitalization, and a total of 7696 patients participated in the observational study. Our model was evaluated against different time windows. The best performance was achieved with an AUC of 0.914 when the observation window was 5 days and the prediction window was 30 days. Outperformed other baseline models including LR (0.708), RF (0.717), SVM (0.675), LSTM (0.757), GRU (0.759), GRU-U (0.766) and MTSSP (0.770). This tool allows us to predict the expected pathway of heart failure patients and intervene early in the treatment process, which has significant implications for improving the life expectancy of heart failure patients.

Machine learning-based prognostic modeling of patients with acute heart failure receiving furosemide in intensive care units

Purpose

This study developed machine learning models to predict in-hospital mortality, initiation of acute renal replacement therapy, and mechanical ventilation in patients with acute heart failure receiving furosemide in intensive care units.

Method

An extensive database comprising static and dynamic features obtained from a Japanese hospital chain was used to construct and train the machine learning models.

Results

The results revealed that the proposed machine learning models predict in-hospital mortality, initiation of acute renal replacement therapy, and mechanical ventilation with good accuracy. However, the optimal models vary depending on the predicted outcomes. The linear support vector machine classification models exhibited the highest in-hospital mortality and mechanical ventilation prediction accuracy, with the area under the receiver operating characteristic curve of 0.73 and 0.73, respectively, whereas the multi-layer neural network exhibited the highest accuracy for acute renal replacement therapy initiation prediction with an area under the receiver operating characteristic curve of 0.70.

Conclusions

In conclusion, this study demonstrated that machine learning models could help predict the clinical outcomes of patients with acute heart failure receiving furosemide. However, the optimal models may differ depending on the outcome of interest.

Learning and visualizing chronic latent representations using electronic health records

Background

Nowadays, patients with chronic diseases such as diabetes and hypertension have reached alarming numbers worldwide. These diseases increase the risk of developing acute complications and involve a substantial economic burden and demand for health resources. The widespread adoption of Electronic Health Records (EHRs) is opening great opportunities for supporting decision-making. Nevertheless, data extracted from EHRs are complex (heterogeneous, high-dimensional and usually noisy), hampering the knowledge extraction with conventional approaches.

Methods

We propose the use of the Denoising Autoencoder (DAE), a Machine Learning (ML) technique allowing to transform high-dimensional data into latent representations (LRs), thus addressing the main challenges with clinical data. We explore in this work how the combination of LRs with a visualization method can be used to map the patient data in a two-dimensional space, gaining knowledge about the distribution of patients with different chronic conditions. Furthermore, this representation can be also used to characterize the patient’s health status evolution, which is of paramount importance in the clinical setting.

Results

To obtain clinical LRs, we considered real-world data extracted from EHRs linked to the University Hospital of Fuenlabrada in Spain. Experimental results showed the great potential of DAEs to identify patients with clinical patterns linked to hypertension, diabetes and multimorbidity. The procedure allowed us to find patients with the same main chronic disease but different clinical characteristics. Thus, we identified two kinds of diabetic patients with differences in their drug therapy (insulin and non-insulin dependant), and also a group of women affected by hypertension and gestational diabetes. We also present a proof of concept for mapping the health status evolution of synthetic patients when considering the most significant diagnoses and drugs associated with chronic patients.

Conclusion

Our results highlighted the value of ML techniques to extract clinical knowledge, supporting the identification of patients with certain chronic conditions. Furthermore, the patient’s health status progression on the two-dimensional space might be used as a tool for clinicians aiming to characterize health conditions and identify their more relevant clinical codes.

Supplementary Information

The online version contains supplementary material available at (10.1186/s13040-022-00303-z).

Unstructured clinical notes within the 24 hours since admission predict short, mid & long-term mortality in adult ICU patients

Mortality prediction for intensive care unit (ICU) patients is crucial for improving outcomes and efficient utilization of resources. Accessibility of electronic health records (EHR) has enabled data-driven predictive modeling using machine learning. However, very few studies rely solely on unstructured clinical notes from the EHR for mortality prediction. In this work, we propose a framework to predict short, mid, and long-term mortality in adult ICU patients using unstructured clinical notes from the MIMIC III database, natural language processing (NLP), and machine learning (ML) models. Depending on the statistical description of the patients' length of stay, we define the short-term as 48-hour and 4-day period, the mid-term as 7-day and 10-day period, and the long-term as 15-day and 30-day period after admission. We found that by only using clinical notes within the 24 hours of admission, our framework can achieve a high area under the receiver operating characteristics (AU-ROC) score for short, mid and long-term mortality prediction tasks. The test AU-ROC scores are 0.87, 0.83, 0.83, 0.82, 0.82, and 0.82 for 48-hour, 4-day, 7-day, 10-day, 15-day, and 30-day period mortality prediction, respectively. We also provide a comparative study among three types of feature extraction techniques from NLP: frequency-based technique, fixed embedding-based technique, and dynamic embedding-based technique. Lastly, we provide an interpretation of the NLP-based predictive models using feature-importance scores.

The (decision) tree of fertility: an innovative decision-making algorithm in assisted reproduction technique

PURPOSE

Several mathematical models have been developed to estimate individualized chances of assisted reproduction techniques (ART) success, although with limited clinical application. Our study aimed to develop a decisional algorithm able to predict pregnancy and live birth rates after controlled ovarian stimulation (COS) phase, helping the physician to decide whether to perform oocytes pick-up continuing the ongoing ART path.

METHODS

A single-center retrospective analysis of real-world data was carried out including all fresh ART cycles performed in 1998-2020. Baseline characteristics, ART parameters and biochemical/clinical pregnancies and live birth rates were collected. A seven-steps systematic approach for model development, combining linear regression analyses and decision trees (DT), was applied for biochemical, clinical pregnancy, and live birth rates.

RESULTS

Of fresh ART cycles, 12,275 were included. Linear regression analyses highlighted a relationship between number of ovarian follicles > 17 mm detected at ultrasound before pick-up (OF17), embryos number and fertilization rate, and biochemical and clinical pregnancy rates (p < 0.001), but not live birth rate. DT were created for biochemical pregnancy (statistical power-SP:80.8%), clinical pregnancy (SP:85.4%), and live birth (SP:87.2%). Thresholds for OF17 entered in all DT, while sperm motility entered the biochemical pregnancy's model, and female age entered the clinical pregnancy and live birth DT. In case of OF17 < 3, the chance of conceiving was < 6% for all DT.

CONCLUSION

A systematic approach allows to identify OF17, female age, and sperm motility as pre-retrieval predictors of ART outcome, possibly reducing the socio-economic burden of ART failure, allowing the clinician to perform or not the oocytes pick-up.

Multi-Task Learning with Recurrent Neural Networks for ARDS Prediction using only EHR Data: Model Development and Validation Study (Preprint)

Background

Acute respiratory distress syndrome (ARDS) is a condition that is often considered to have broad and subjective diagnostic criteria and is associated with significant mortality and morbidity. Early and accurate prediction of ARDS and related conditions such as hypoxemia and sepsis could allow timely administration of therapies, leading to improved patient outcomes.

Objective

The aim of this study is to perform an exploration of how multilabel classification in the clinical setting can take advantage of the underlying dependencies between ARDS and related conditions to improve early prediction of ARDS in patients.

Methods

The electronic health record data set included 40,703 patient encounters from 7 hospitals from April 20, 2018, to March 17, 2021. A recurrent neural network (RNN) was trained using data from 5 hospitals, and external validation was conducted on data from 2 hospitals. In addition to ARDS, 12 target labels for related conditions such as sepsis, hypoxemia, and COVID-19 were used to train the model to classify a total of 13 outputs. As a comparator, XGBoost models were developed for each of the 13 target labels. Model performance was assessed using the area under the receiver operating characteristic curve. Heat maps to visualize attention scores were generated to provide interpretability to the neural networks. Finally, cluster analysis was performed to identify potential phenotypic subgroups of patients with ARDS.

Results

The single RNN model trained to classify 13 outputs outperformed the individual XGBoost models for ARDS prediction, achieving an area under the receiver operating characteristic curve of 0.842 on the external test sets. Models trained on an increasing number of tasks resulted in improved performance. Earlier prediction of ARDS nearly doubled the rate of in-hospital survival. Cluster analysis revealed distinct ARDS subgroups, some of which had similar mortality rates but different clinical presentations.

Conclusions

The RNN model presented in this paper can be used as an early warning system to stratify patients who are at risk of developing one of the multiple risk outcomes, hence providing practitioners with the means to take early action.

Usability of Electronic Health records in Predicting Short-term falls: Machine learning Applications in Senior Care Facilities (Preprint)

Background

Short-term fall prediction models that use electronic health records (EHRs) may enable the implementation of dynamic care practices that specifically address changes in individualized fall risk within senior care facilities.

Objective

The aim of this study is to implement machine learning (ML) algorithms that use EHR data to predict a 3-month fall risk in residents from a variety of senior care facilities providing different levels of care.

Methods

This retrospective study obtained EHR data (2007-2021) from Juniper Communities’ proprietary database of 2785 individuals primarily residing in skilled nursing facilities, independent living facilities, and assisted living facilities across the United States. We assessed the performance of 3 ML-based fall prediction models and the Juniper Communities’ fall risk assessment. Additional analyses were conducted to examine how changes in the input features, training data sets, and prediction windows affected the performance of these models.

Results

The Extreme Gradient Boosting model exhibited the highest performance, with an area under the receiver operating characteristic curve of 0.846 (95% CI 0.794-0.894), specificity of 0.848, diagnostic odds ratio of 13.40, and sensitivity of 0.706, while achieving the best trade-off in balancing true positive and negative rates. The number of active medications was the most significant feature associated with fall risk, followed by a resident’s number of active diseases and several variables associated with vital signs, including diastolic blood pressure and changes in weight and respiratory rates. The combination of vital signs with traditional risk factors as input features achieved higher prediction accuracy than using either group of features alone.

Conclusions

This study shows that the Extreme Gradient Boosting technique can use a large number of features from EHR data to make short-term fall predictions with a better performance than that of conventional fall risk assessments and other ML models. The integration of routinely collected EHR data, particularly vital signs, into fall prediction models may generate more accurate fall risk surveillance than models without vital signs. Our data support the use of ML models for dynamic, cost-effective, and automated fall predictions in different types of senior care facilities.