Using machine learning techniques to predict hospital admission at the emergency department (original) (raw)
Related papers
Using Machine Learning for Predicting the Hospitalization of Emergency Department Patients
Studies in Health Technology and Informatics
Artificial intelligence processes are increasingly being used in emergency medicine, notably for supporting clinical decisions and potentially improving healthcare services. This study investigated demographics, coagulation tests, and biochemical markers routinely used for patients seen in the Emergency Department (ED) concerning hospitalization. This retrospective observational study included 13,991 emergency department visits of patients who had undergone biomarker testing to a tertiary public hospital in Greece during 2020. After applying five well-known classifiers of the caret package for machine learning of the R programming language in the whole data set and to each ED unit separately, the best performance regarding AUC ROC was observed in the Pulmonology ED unit. Furthermore, among the five classification techniques evaluated, a random forest classifier outperformed other models.
Predicting Hospital Admission for Emergency Department Patients: A Machine Learning Approach
Studies in Health Technology and Informatics
The objective of this study was to establish a machine learning model and to evaluate its predictive capability of admission to the hospital. This observational retrospective study included 3204 emergency department visits to a public tertiary care hospital in Greece from 14 March to 4 May 2019. We investigated biochemical markers and coagulation tests that are routinely checked in patients visiting the Emergency Department (ED) in relation to the ED outcome (admission or discharge). Among the most popular classification techniques of the scikit-learn library through a 10-fold cross-validation approach, a GaussianNB model outperformed other models with respect to the area under the receiver operating characteristic curve.
MACHINE LEARNING TECHNIQUES TO PREDICT THE HOSPITAL ADMISSIONS FROM EMERGENCY DEPARTMENTS
IAEME PUBLICATION, 2020
Healthcare companies frequently profit from information technologies as well as enclosed decision support methods, which enhance the quality of services (QoS) and support to limit difficulties and conflicting effects. To be prepared to predict, at the time of triage, whether a requirement for hospital access endures for emergency department (ED) patients may discover valuable data that could provide to system-wide hospital improvements intended to advance ED throughput. The main objective of this study is to improve and prove a predictive model to evaluate whether a patient is expected to need inpatient access at the time of ED triage, using regular hospital official data. The Machine Learning schemes for healthcare in developing algorithms applied to recognize the complex patterns with big data. In this paper, we use three machine learning algorithms to improve the predictive models: (1) logistic regression (LR), (2) Modified Random Forest (MRF), and (3) gradient boosted machines (GBM). The GBM performed better (accuracy=82.41%) than the modified random forest (accuracy=81.09%) and the logistic regression model (accuracy=80.54%). Drawing on logistic regression, we recognize various hospital admissions circumstances, including hospital site, care group, age, arrival mode, attributes category, the last admission in the past month, and earlier access in the past year. This research highlights the possible benefit of three standard machine learning algorithms in predicting patient admissions.
Predicting Emergency Department Services in Hospital Using Machine Learning
2019
Healthcare organizations often benefit from information technologies as well as embedded decision support systems, which improve the quality of services and help preventing complications and adverse events. To be able to predict, at the time of triage, whether a need for hospital admission exists for emergency department (ED) patients may constitute useful information that could contribute to system wide hospital changes designed to improve ED throughput. The objective of this study was to develop and validate a predictive model to assess whether a patient is likely to require inpatient admission at the time of ED triage, using routine hospital administrative data. Using Single Data Mining Technique hospital admissions for the emergency department has been comprehensively investigated showing acceptable levels of accuracy. The Machine Learning technique for Healthcare in developing algorithms that is used to identify the complex patterns with large amount of data. This technique imp...
Using Data Mining to Predict Hospital Admissions from the Emergency Department
Crowding within emergency departments can have significant negative consequences for patients. EDs therefore need to explore the use of innovative methods to improve patient flow and prevent overcrowding. One potential method is the use of data mining using machine learning techniques to predict ED admissions. This paper uses routinely collected administrative data (120 600 records) from two major acute hospitals in Northern Ireland to compare contrasting machine learning algorithms in predicting the risk of admission from the ED. We use three algorithms to build the predictive models: 1) logistic regression; 2) decision trees; and 3) gradient boosted machines (GBM). The GBM performed better (accuracy D 80:31%, AUC-ROC D 0:859) than the decision tree (accuracy D 80:06%, AUC-ROC D 0:824) and the logistic regression model (accuracy D 79:94%, AUC-ROC D 0:849). Drawing on logistic regression, we identify several factors related to hospital admissions, including hospital site, age, arrival mode, triage category, care group, previous admission in the past month, and previous admission in the past year. Decision support systems may be able to offer a picture of expected ED admissions at any given moment as a consequence of this study, allowing for resource planning ahead of time and avoiding patient flow bottlenecks. This research also suggests that the models described in this study may be utilised to perform comparisons between projected and actual admission rates. Generalised bivariate models (GBMs) are sufficient when interpretability is a concern; however, if accuracy is crucial, logistic regression models should be considered. INDEX TERMS{IT} Datamining,emergency department, hospitals, machine learning, and predictive models.
Analysis on Benefits and Costs of Machine Learning-Based Early Hospitalization Prediction
IEEE Access
Overcrowding in emergency departments (EDs) has long been a problem worldwide and has serious consequences for patient satisfaction and safety. Typically, overcrowding is caused by delays in the boarding time of ED patients waiting for inpatient beds. If the hospitalization of patients is predicted early enough in EDs, inpatient beds can be prepared in advance and the boarding time can be reduced. We design machine learning-based hospitalization predictive models using data on 27,747 patients and compare the experimental results. Five predictive models are designed: 1) logistic regression, 2) XGBoost, 3) NGBoost, 4) support vector machine, and 5) decision tree models. Based on the predictive results, we estimate the quantitative effects of hospitalization predictions on EDs and wards. Using the data from the ED of a general hospital in South Korea, our experiments show that the ED length of stay of a patient can be reduced by 12.3 minutes on average and the ED can reduce the total length of stay by 340,147 minutes for a year. INDEX TERMS Emergency department, machine learning, hospitalization prediction, estimation of quantitative effects.
Machine Learning Approaches to Predict Patient’s Length of Stay in Emergency Department
Applied Computational Intelligence and Soft Computing
As the COVID-19 pandemic has affected the globe, health systems worldwide have also been significantly affected. This pandemic has impacted many sectors, including health in the Kingdom of Jordan. Crises that put heavy pressure on the health systems’ shoulders include the emergency departments (ED), the most demanded hospital resources during normal conditions, and critical during crises. However, managing the health systems efficiently and achieving the best planning and allocation of their EDs’ resources becomes crucial to improve their capabilities to accommodate the crisis’s impact. Knowing critical factors affecting the patient length of stay prediction is critical to reducing the risks of prolonged waiting and clustering inside EDs. That is, by focusing on these factors and analyzing the effect of each. This research aims to determine the critical factors that predict the outcome: the length of stay, i.e., the predictor variables. Therefore, patients’ length of stay in EDs acr...
Predictive models for emergency department triage using machine learning a review
Background: Recently, many research groups have tried to develop emergency department triage decision support systems based on big volumes of historical clinical data to differentiate and prioritize patients. Machine learning models might improve the predictive capacity of emergency department triage systems. The aim of this review was to assess the performance of recently described machine learning models for patient triage in emergency departments, and to identify future challenges. Methods: Four databases (ScienceDirect, PubMed, Google Scholar and Springer) were searched using key words identified in the research questions. To focus on the latest studies on the subject, the most cited papers between 2018 and October 2021 were selected. Only works with hospital admission and critical illness as outcomes were included in the analysis. Results: Eleven articles concerned the two outcomes (hospital admission and critical illness) and developed 55 predictive models. Random Forest and Logistic Regression were the most commonly used prediction algorithms, and the receiver operating characteristic-area under the curve (ROC-AUC) the most frequently used metric to assess the algorithm prediction performance. Random Forest and Logistic Regression were the most discriminant models according to the selected studies. Conclusions: Machine learning-based triage systems could improve decision-making in emergency depart-ments, thus leading to better patients’ outcomes. However, there is still scope for improvement concerning the prediction performance and explica-bility of ML models.
European Journal of Medical Research, 2022
Sepsis is an inflammation caused by the body's systemic response to an infection. The infection could be a result of many diseases, such as pneumonia, urinary tract infection, and other illnesses. Some of its symptoms are fever, tachycardia, tachypnea, etc. Unfortunately, sepsis remains a critical problem at the hospitals and leads to many issues, such as increasing mortality rate, health care costs, and health care utilization. Early detection of sepsis in patients can help respond quickly, take preventive actions, and prevent major issues. The main aim of this study is to predict the risk of sepsis by utilizing the patient's demographic and clinical information, i.e., patient's gender, age, severity level, mortality risk, admit type along with hospital length of stay. Six machine learning approaches, Logistic Regression (LR), Naïve Bayes, Support Vector Machine (SVM), Boosted Tree, Classification and Regression Tree (CART), and Bootstrap Forest are used to predict the risk of sepsis. The results showed that different machine learning methods have other performances in terms of various measures. For instance, the Bootstrap Forest machine learning method exhibited the highest performance in AUC and R-square or SVM and Boosted Tree showed the highest performance in terms of misclassification rate. The Bootstrap Forest can be considered the best machine learning method in predicting sepsis regarding applied features in this research, mainly because it showed superior performance and efficiency in two performance measures: AUC and R-square. Highlights • Six machine learning methods, Logistic Regression (LR), Naïve Bayes, Support Vector Machine (SVM), Boosted Tree, Classification and Regression Tree (CART), and Bootstrap Forest were compared together in order to predict sepsis. • Early stage of admission data including patient's gender, age, severity level, mortality risk, admit type along with hospital length of stay were used for predicting sepsis. • The Bootstrap Forest can be considered the best machine learning method in predicting sepsis regarding applied features in this research mainly because it showed superior performance and efficiency in two performance measures, i.e. AUC and R-square.
PLOS ONE
The risk stratification of patients in the emergency department begins at triage. It is vital to stratify patients early based on their severity, since undertriage can lead to increased morbidity, mortality and costs. Our aim was to present a new approach to assist healthcare professionals at triage in the stratification of patients and in identifying those with higher risk of ICU admission. Adult patients assigned Manchester Triage System (MTS) or Emergency Severity Index (ESI) 1 to 3 from a Portuguese and a United States Emergency Departments were analyzed. Variables routinely collected at triage were used and natural language processing was applied to the patient chief complaint. Stratified random sampling was applied to split the data in train (70%) and test (30%) sets and 10-fold cross validation was performed for model training. Logistic regression, random forests, and a random undersampling boosting algorithm were used. We compared the performance obtained with the reference model-using only triage priorities-with the models using additional variables. For both hospitals, a logistic regression model achieved higher overall performance, yielding areas under the receiver operating characteristic and precision-recall curves of 0.91 (95% CI 0.90-0.92) and 0.30 (95% CI 0.27-0.33) for the United States hospital and of 0.85 (95% CI 0.83-0.86) and 0.06 (95% CI 0.05-0.07) for the Portuguese hospital. Heart rate, pulse oximetry, respiratory rate and systolic blood pressure were the most important predictors of ICU admission. Compared to the reference models, the models using clinical variables and the chief complaint presented higher recall for patients assigned MTS/ESI 3 and can identify patients assigned MTS/ESI 3 who are at risk for ICU admission.