Predictive Modeling for COVID-19 Emergency Department Stay using Machine Learning

  • Unique Paper ID: 167496
  • Volume: 11
  • Issue: 3
  • PageNo: 1225-1230
  • Abstract:
  • The COVID-19 pandemic has significantly increased emergency department (ED) stays for patients in the United States. To address this issue, a study aimed to create a reliable model predicting the length of stay (LOS) for COVID-19 patients in the ED and identify factors influencing meeting the '4-hour target.' Data from diverse urban hospitals in Detroit, collected from March 16 to December 29, 2020, informed this research. Using data processing, four machine learning models (logistic regression, gradient boosting, decision tree, and random forest) were trained to forecast whether COVID-19 patients' ED stays would surpass 4 hours. The study involved 3,301 patients with 16 clinical factors. The gradient boosting (GB) model outperformed others, achieving 85% accuracy and an F1-score of 0.88 in predicting LOS within the test data, surpassing the logistic regression baseline, decision tree, and random forest models. Further data splitting did not notably improve accuracy. This investigation identified critical factors, including patient demographics, existing health conditions, and operational ED data, as predictors of extended stays for COVID-19 patients. The predictive model could serve as a decision-making tool, aiding in resource planning for EDs and hospitals. Moreover, it offers patients estimations of their ED LOS, enhancing their understanding and potentially improving their experience. In summary, this study's model effectively predicts ED LOS for COVID-19 patients, enabling better resource allocation and informed decision-making in managing ED stays during the pandemic, potentially improving patient care and hospital efficiency.

Cite This Article

  • ISSN: 2349-6002
  • Volume: 11
  • Issue: 3
  • PageNo: 1225-1230

Predictive Modeling for COVID-19 Emergency Department Stay using Machine Learning

Related Articles