Forecasting emergency department arrivals: a tutorial for emergency department directors

Area of Science:

• Health Services Research
• Operations Research
• Biostatistics

Background:

• Effective strategic, tactical, and operational planning in emergency departments (EDs) requires accurate forecasting of patient arrivals.
• Existing forecasting methods may not fully capture the complex temporal variations in ED demand.

Purpose of the Study:

• To present and demonstrate regression-based forecasting models for predicting emergency department (ED) arrivals.
• To provide a practical tool for ED medical directors to support resource planning activities.

Main Methods:

• Utilized trend regression analysis on annual ED arrival data to identify long-term growth.
• Employed zero/one variables to model monthly and daily variations in ED arrivals.
• Applied Fourier regression to capture hourly, cyclical patterns in ED patient flow.

Main Results:

• Identified an average annual increase in ED demand of approximately 1,000 arrivals.
• Determined July as the busiest month and February as the slowest.
• Observed approximately 20 fewer arrivals on Fridays compared to Sundays.
• Found peak ED arrivals between 1 p.m. and 6 p.m., averaging 10 arrivals per hour.

Conclusions:

• Regression-based forecasting models are versatile and can accommodate various ED arrival prediction scenarios.
• These models offer valuable insights into ED demand patterns, supporting informed resource allocation.
• The widespread availability of hospital data and regression software facilitates the adoption of these forecasting tools for ED management.