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

Hsu, Chih‐Yi; Chu, Cheng-CJ; Lin, Ching-Heng; Huang, Chia-Jung; Ng, Chip-Jin; Lin, Guan‐Yu; Chiou, Meng-Jiun; Lo, Huey‐Ming; Chen, Shou-Yen

doi:10.3390/diagnostics12010082

Cited by 11 publications

(15 citation statements)

References 44 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Also, our report demonstrated that ML would be a better approach to provide a prediction model than multivariate logistic regression, which mainly focuses on the association between dependent and independent variables. The AUROC in this study was almost the same as that in previous studies that used ML technique, approximately 0.74–0.83 with a different algorithm 13–15. However, in some studies with multivariate logistic regression, the concordance (C)-statistic, which is often used to assess the ability of a risk factor to predict outcome, ranged from 0.55 to 0.74 4.…”

Section: Discussionsupporting

confidence: 76%

“…Hong et al indicated that gradient-boosting models that leveraged clinical data were superior to traditional logistic regression models built on administrative data to predict ED revisit 14. Hsu et al developed an ML model, the voting classifier model, to predict ED revisit in patients with abdominal pain 15. These works shed light on the use of a prediction model for ED revisit based on an ML algorithm.…”

Section: Introductionmentioning

confidence: 99%

“… 14 Hsu et al developed an ML model, the voting classifier model, to predict ED revisit in patients with abdominal pain. 15 These works shed light on the use of a prediction model for ED revisit based on an ML algorithm.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Prediction of high-risk emergency department revisits from a machine-learning algorithm: a proof-of-concept study

Sung,

Ho,

Fan

et al. 2024

BMJ Health Care Inform

View full text Add to dashboard Cite

BackgroundHigh-risk emergency department (ED) revisit is considered an important quality indicator that may reflect an increase in complications and medical burden. However, because of its multidimensional and highly complex nature, this factor has not been comprehensively investigated. This study aimed to predict high-risk ED revisit with a machine-learning (ML) approach.MethodsThis 3-year retrospective cohort study assessed adult patients between January 2019 and December 2021 from National Taiwan University Hospital Hsin-Chu Branch with high-risk ED revisit, defined as hospital or intensive care unit admission after ED return within 72 hours. A total of 150 features were preliminarily screened, and 79 were used in the prediction model. Deep learning, random forest, extreme gradient boosting (XGBoost) and stacked ensemble algorithm were used. The stacked ensemble model combined multiple ML models and performed model stacking as a meta-level algorithm. Confusion matrix, accuracy, sensitivity, specificity and area under the receiver operating characteristic curve (AUROC) were used to evaluate performance.ResultsAnalysis was performed for 6282 eligible adult patients: 5025 (80.0%) in the training set and 1257 (20.0%) in the testing set. High-risk ED revisit occurred for 971 (19.3%) of training set patients vs 252 (20.1%) in the testing set. Leading predictors of high-risk ED revisit were age, systolic blood pressure and heart rate. The stacked ensemble model showed more favourable prediction performance (AUROC 0.82) than the other models: deep learning (0.69), random forest (0.78) and XGBoost (0.79). Also, the stacked ensemble model achieved favourable accuracy and specificity.ConclusionThe stacked ensemble algorithm exhibited better prediction performance in which the predictions were generated from different ML algorithms to optimally maximise the final set of results. Patients with older age and abnormal systolic blood pressure and heart rate at the index ED visit were vulnerable to high-risk ED revisit. Further studies should be conducted to externally validate the model.

show abstract

Section: Discussionsupporting

confidence: 76%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Prediction of high-risk emergency department revisits from a machine-learning algorithm: a proof-of-concept study

Sung,

Ho,

Fan

et al. 2024

BMJ Health Care Inform

View full text Add to dashboard Cite

show abstract

“…The 72-hour unscheduled ED revisit rate is considered a healthcare quality indicator, as unscheduled ED revisits are associated with ED crowding, higher healthcare expenditure, and poorer medical experiences (28-30). Previous studies have also revealed that mental illness is a risk factor for unscheduled ED revisits and have attributed this to unclear discharge instructions, ine cient discharge systems, and inadequate post-ED follow-ups (31)(32)(33). To address this issue, various strategies such as multidisciplinary approaches, integrated discharge systems, and post-ED care programs have been found to be effective in reducing unscheduled ED revisits (34,35).…”

Section: Discussionmentioning

confidence: 99%

Severe Mental Illness and Non-specific Abdominal Pain in the Emergency Department: A Multi-institutional Database Study

Chien

Kou

et al. 2023

Preprint

Self Cite

View full text Add to dashboard Cite

Introduction: Abdominal pain is a common cause of emergency department (ED) visits, and non-specific abdominal pain (NSAP) accounts for a large proportion of diagnoses. Patients with severe mental illness (SMI) are particularly vulnerable due to their atypical disease presentation and high comorbidity rates. Previous studies have reported higher ED revisit rates and delayed diagnoses in patients with SMI and NSAP. This study aimed to evaluate ED management, unscheduled ED revisit rates, and short-term adverse outcomes in patients with SMI and NSAP.Methods: Relying on the Chang Gung Research Database (CGRD), we selectively used data from January 1, 2007, to December 31, 2017. Diagnoses of NSAP and SMI were confirmed by combining the ICD codes with relevant medical records. The non-SMI group was matched at a ratio of 1:3 using a Greedy algorithm. The outcomes were ED management, 72-hour unscheduled ED revisits, and 7-day adverse events.Results: A total of 233,671 patients from seven hospitals over a span of 11 years were recruited; among them, 98,722 were excluded based on the inclusion criteria, leaving 134,949 patients for analysis. The SMI group had more comorbidities, a higher rate of 72-hour unscheduled ED revisits, and was more likely to receive analgesics, but less likely to undergo laboratory tests and CT scans. Patients without SMI were more likely to be admitted to the ward and undergo invasive abdominal procedures within seven days after index ED discharge. No significant differences were found in ICU admission, abdominal surgery, or in-hospital mortality between the two groups.Conclusion: Our study demonstrated that patients with SMI and NSAP had a higher rate of 72-hour unscheduled ED revisits, but this did not lead to higher short-term adverse outcomes. Although NSAP is considered a safe diagnosis for both the general population and patients with SMI, the higher rate of unscheduled ED revisits suggests the need for better healthcare interventions to eliminate health disparities in this vulnerable group.

show abstract

“…Previously, we attempted to predict 72-hour URVs in ED patients with abdominal pain using the ML model. [22] A crucial yet unfulfilled objective is to validate this model using future data. In this study, we not only refined our data to adjust our previous prediction model but also evaluated the model performance in future data validation during the COVID-19 era.…”

Section: Introductionmentioning

confidence: 99%

Machine learning models for predicting unscheduled return visits of patients with abdominal pain at emergency department and validation during COVID-19 pandemic: A retrospective cohort study

Hsu,

Chu,

et al. 2024

Medicine

Self Cite

View full text Add to dashboard Cite

Machine learning (ML) models for predicting 72-hour unscheduled return visits (URVs) for patients with abdominal pain in the emergency department (ED) were developed in a previous study. This study refined the data to adjust previous prediction models and evaluated the model performance in future data validation during the COVID-19 era. We aimed to evaluate the practicality of the ML models and compare the URVs before and during the COVID-19 pandemic. We used electronic health records from Chang Gung Memorial Hospital from 2018 to 2019 as a training dataset, and various machine learning models, including logistic regression (LR), random forest (RF), extreme gradient boosting (XGB), and voting classifier (VC) were developed and subsequently used to validate against the 2020 to 2021 data. The models highlighted several determinants for 72-hour URVs, including patient age, prior ER visits, specific vital signs, and medical interventions. The LR, XGB, and VC models exhibited the same AUC of 0.71 in the testing set, whereas the VC model displayed a higher F1 score (0.21). The XGB model demonstrated the highest specificity (0.99) and precision (0.64) but the lowest sensitivity (0.01). Among these models, the VC model showed the most favorable, balanced, and comprehensive performance. Despite the promising results, the study illuminated challenges in predictive modeling, such as the unforeseen influences of global events, such as the COVID-19 pandemic. These findings not only highlight the significant potential of machine learning in augmenting emergency care but also underline the importance of iterative refinement in response to changing real-world conditions.

show abstract

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

Cited by 11 publications

References 44 publications

Prediction of high-risk emergency department revisits from a machine-learning algorithm: a proof-of-concept study

Prediction of high-risk emergency department revisits from a machine-learning algorithm: a proof-of-concept study

Severe Mental Illness and Non-specific Abdominal Pain in the Emergency Department: A Multi-institutional Database Study

Machine learning models for predicting unscheduled return visits of patients with abdominal pain at emergency department and validation during COVID-19 pandemic: A retrospective cohort study

Contact Info

Product

Resources

About