An Integrated Forecasting Prototype for Emergency Department Boarding Time to Support Proactive Operational Decision Making

2026-05-12

AI Summary

This study addresses the challenge of emergency department (ED) boarding caused by inpatient bed shortages, which exacerbates healthcare congestion and delays in patient care. The authors propose a multi-step time series forecasting framework that, for the first time, applies DLinear and NLinear models to predict ED boarding duration over horizons of 6 to 24 hours. The approach integrates external contextual features—including weather conditions, holidays, and local major events—to enhance predictive accuracy. An MLOps prototype system is developed to support real-time operational decision-making through automated data ingestion, visualization, and model retraining. Experimental results demonstrate that the models achieve strong performance across multiple forecast horizons and maintain robustness even under extreme congestion scenarios, thereby enabling proactive resource allocation and timely interventions.

Abstract

Overcrowding in emergency departments (ED) remains a persistent operational challenge worldwide, causing delays in care delivery and downstream congestion. ED boarding time, defined as the duration admitted patients remain in the ED while awaiting inpatient bed placement, is a key indicator of this congestion. Predicting ED boarding time in advance enables proactive operational decision making before congestion escalates. We developed and evaluated a multi-horizon time series forecasting framework to predict ED boarding time at 6, 8, 10, 12, and 24-hour horizons. Real-world data from a university-affiliated urban hospital in the United States were utilized and integrated with external contextual data sources, including weather, holidays, and major local events. Decomposition-based Linear (DLinear) and Normalization-based Linear (NLinear) time series forecasting deep learning models showed superior performance across multiple horizons. Models were also evaluated under extreme congestion scenarios characterized by elevated boarding times. In addition, a Machine Learning Operations (MLOps) web application prototype was developed to support translation of the forecasting framework into practice through integrated data ingestion, forecast visualization, experimentation, and retraining.

Problem

• emergency department overcrowding
• boarding time
• congestion prediction
• proactive decision making
• operational challenge

Innovation

• multi-horizon forecasting
• DLinear
• NLinear
• MLOps
• emergency department boarding time