Predicting 72-hour emergency department revisits

Pellerin, Gene; Gao, Kelly; Kaminsky, Laurence S.

doi:10.1016/j.ajem.2017.08.049

Cited by 28 publications

(23 citation statements)

References 25 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…To our knowledge, this study is the first to examine the use of machine learning in the prediction of 72 h URVs for patients with abdominal pain presenting to the ED. Previous studies have reported prediction models using machine learning to predict 72 h URV for all causes, but no previous studies have focused on ED patients with abdominal pain [ 18 , 19 , 24 ]. One study examined the relationship between triage level and return visits in patients with abdominal pain, but the study lacked a prediction model for these patients [ 25 ].…”

Section: Discussionmentioning

confidence: 99%

“…A greater number of X-rays taken in a given time period may imply that the patient’s condition is more complicated and that the patient probably has other underlying conditions. For these patients with comorbidities, dispositions should be made carefully after thorough deliberation [ 19 , 41 ]. According to our study, close observation of the symptoms or the arranging of further exams should be considered for these patients.…”

Section: Discussionmentioning

confidence: 99%

“…Because of these characteristics, machine learning can assist physicians in multiple clinical situations, such as image study classification and mortality prediction [ 16 , 17 ]. Previous studies have attempted to use machine learning models to predict the 72 h URVs of ED patients [ 18 , 19 ]. These previous models attempted to predict URVs for all causes, which is a highly complex task and may have reduced the accuracy of prediction.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

A Machine Learning Model for Predicting Unscheduled 72 h Return Visits to the Emergency Department by Patients with Abdominal Pain

et al. 2021

View full text Add to dashboard Cite

Seventy-two-hour unscheduled return visits (URVs) by emergency department patients are a key clinical index for evaluating the quality of care in emergency departments (EDs). This study aimed to develop a machine learning model to predict 72 h URVs for ED patients with abdominal pain. Electronic health records data were collected from the Chang Gung Research Database (CGRD) for 25,151 ED visits by patients with abdominal pain and a total of 617 features were used for analysis. We used supervised machine learning models, namely logistic regression (LR), support vector machine (SVM), random forest (RF), extreme gradient boosting (XGB), and voting classifier (VC), to predict URVs. The VC model achieved more favorable overall performance than other models (AUROC: 0.74; 95% confidence interval (CI), 0.69–0.76; sensitivity, 0.39; specificity, 0.89; F1 score, 0.25). The reduced VC model achieved comparable performance (AUROC: 0.72; 95% CI, 0.69–0.74) to the full models using all clinical features. The VC model exhibited the most favorable performance in predicting 72 h URVs for patients with abdominal pain, both for all-features and reduced-features models. Application of the VC model in the clinical setting after validation may help physicians to make accurate decisions and decrease URVs.

show abstract

Section: Discussionmentioning

confidence: 99%

Section: Discussionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

A Machine Learning Model for Predicting Unscheduled 72 h Return Visits to the Emergency Department by Patients with Abdominal Pain

et al. 2021

View full text Add to dashboard Cite

show abstract

“…The characteristics of general ED revisits have been studied, and prediction models to identify general revisits have been developed. [23][24][25] Prediction models of high-risk revisits are quite limited, as such models would require a large sample size to predict rare events. Prediction can occur at initial ED discharge (most common), between visits, or upon the revisit.…”

Section: Discussionmentioning

confidence: 99%

High-risk Return Visits to United States Emergency Departments, 2010–2018

Ling

Sung

Fang

et al. 2022

WestJEM

View full text Add to dashboard Cite

Introduction: Although factors related to a return visit to the emergency department (ED) have been reported, only a few studies have examined “high-risk” ED revisits with serious adverse outcomes. In this study we aimed to describe the incidence and trend of high-risk ED revisits in United States EDs and to investigate factors associated with these revisits. Methods: We obtained data from the National Hospital Ambulatory Medical Care Survey (NHAMCS), 2010–2018. Adult ED revisits within 72 hours of a previous discharge were identified using a mark on the patient record form. We defined high-risk revisits as revisits with serious adverse outcomes, including intensive care unit admissions, emergency surgery, cardiac catheterization, or cardiopulmonary resuscitation (CPR) during the return visit. We performed analyses using descriptive statistics and multivariable logistic regression, accounting for NHAMCS’s complex survey design. Results: Over the nine-year study period, there were an estimated 37,700,000 revisits, and the proportion of revisits in the entire ED population decreased slightly from 5.1% in 2010 to 4.5% in 2018 (P for trend = 0.02). By contrast, there were an estimated 827,000 high-risk ED revisits, and the proportion of high-risk revisits in the entire ED population remained stable at approximately 0.1%. The mean age of these high-risk revisit patients was 57 years, and 43% were men. Approximately 6% of the patients were intubated, and 13% received CPR. Most of them were hospitalized, and 2% died in the ED. Multivariable analysis showed that older age (65+ years), Hispanic ethnicity, daytime visits, and arrival by ambulance during the revisit were independent predictors of high-risk revisits. Conclusion: High-risk revisits accounted for a relatively small fraction (0.1%) of ED visits. Over the period of the NHAMCS survey between 2010-2018, this fraction remained stable. We identified factors during the return visit that could be used to label high-risk revisits for timely intervention.

show abstract

“…In general, the studies carried out [31] are based on the use of descriptive methods on demographic variables (degree of disability or life situation) or quantitative variables (drug count, time markers, or diagnostic codes). In machine learning, [32] random forests and gradient boosting have been used to predict return within 30 days [33], or logical regression [34] for time intervals shorter than 72 h, such as in [35]. Another study [36] used a gradient boosting over a range of 72 h to nine days to analyze data from electronic clinical records [37] such as administrative data (demographics, previous hospital use, comorbidity categories, historical vital values and current), treatment data (laboratory values, ECG and imaging counts, drugs administered), data available at the time of triage and data available at the time of discharge.…”

Section: Introductionmentioning

confidence: 99%

Application of Machine Learning Techniques to Analyze Patient Returns to the Emergency Department

Cabezuelo

2020

JPM

View full text Add to dashboard Cite

The study of the quality of hospital emergency services is based on analyzing a set of indicators such as the average time of first medical attention, the average time spent in the emergency department, degree of completion of the medical report and others. In this paper, an analysis is presented of one of the quality indicators: the rate of return of patients to the emergency service less than 72 h from their discharge. The objective of the analysis was to know the variables that influence the rate of return and which prediction model is the best. In order to do this, the data of the activity of the emergency service of a hospital of a reference population of 290,000 inhabitants were analyzed, and prediction models were created for the binary objective variable (rate of return to emergencies) using the logistic regression techniques, neural networks, random forest, gradient boosting and assembly models. Each of the models was analyzed and the result shows that the best model is achieved through a neural network with activation function tanh, algorithm levmar and three nodes in the hidden layer. This model obtains the lowest mean squared error (MSE) and the best area under the curve (AUC) with respect to the rest of the models used.

show abstract

Predicting 72-hour emergency department revisits

Cited by 28 publications

References 25 publications

A Machine Learning Model for Predicting Unscheduled 72 h Return Visits to the Emergency Department by Patients with Abdominal Pain

A Machine Learning Model for Predicting Unscheduled 72 h Return Visits to the Emergency Department by Patients with Abdominal Pain

High-risk Return Visits to United States Emergency Departments, 2010–2018

Application of Machine Learning Techniques to Analyze Patient Returns to the Emergency Department

Contact Info

Product

Resources

About