Journal of Clinical and Translational Science (Apr 2022)

166 Predicting 30 Day Return Hospital Admissions in Patients with COVID-19 Discharged from the Emergency Department: A national retrospective cohort study

  • David Beiser,
  • Zach Jarou,
  • Michael Puskarich,
  • Marie Vrablik,
  • Elizabeth Rosenman,
  • Samuel McDonald,
  • Jeffrey Kline

DOI
https://doi.org/10.1017/cts.2022.74
Journal volume & issue
Vol. 6
pp. 19 – 19

Abstract

Read online

OBJECTIVES/GOALS: Identification of COVID-19 patients at risk for deterioration following discharge from the emergency department (ED) remains a clinical challenge. Our objective was to develop a prediction model that identifies COVID-19 patients at risk for return and hospital admission within 30 days of ED discharge. METHODS/STUDY POPULATION: We performed a retrospective cohort study of discharged adult ED patients (n = 7,529) with SARS-CoV-2 infection from 116 unique hospitals contributing to the national REgistry of suspected COVID-19 in EmeRgency care (RECOVER). The primary outcome was return hospital admission within 30 days. Models were developed using Classification and Regression Tree (CART), Gradient Boosted Machine (GBM), Random Forest (RF), and least absolute shrinkage and selection (LASSO) approaches. RESULTS/ANTICIPATED RESULTS: Among COVID-19 patients discharged from the ED on their index encounter, 571 (7.6%) returned for hospital admission within 30 days. The machine learning (ML) models (GBM, RF,: and LASSO) performed similarly. The RF model yielded a test AUC of 0.74 (95% confidence interval [CI] 0.71–0.78) with a sensitivity of 0.46 (0.39-0.54) and specificity of 0.84 (0.82-0.85). Predictive variables including: lowest oxygen saturation, temperature; or history of hypertension,: diabetes, hyperlipidemia, or obesity, were common to all ML models. DISCUSSION/SIGNIFICANCE: A predictive model identifying adult ED patients with COVID-19 at risk for return hospital admission within 30 days is feasible. Ensemble/boot-strapped classification methods outperform the single tree CART method. Future efforts may focus on the application of ML models in the hospital setting to optimize allocation of follow up resources.