Forecasting ward-level bed requirements to aid pandemic resource planning: Lessons learned and future directions

Area of Science:

• Health Services Research
• Epidemiology
• Health Informatics

Background:

• COVID-19 pandemic highlighted the need for effective hospital resource management.
• Existing research focused on regional/country-level forecasting, leaving a gap in ward-level planning tools.
• Hospital staff often face challenges in accurately predicting resource needs during health crises.

Purpose of the Study:

• To assess, validate, and deploy a ward-level forecasting tool for hospital resource planning during the COVID-19 pandemic.
• To compare the accuracy of statistical and machine learning forecasting methods for predicting hospital bed occupancy.
• To provide a practical tool for hospital staff to improve pandemic preparedness and patient care.

Main Methods:

• Developed and validated a prototype forecasting tool integrated with a modified Traffic Control Bundling (TCB) protocol.
• Compared statistical and machine learning (ML) forecasting models at two Canadian hospitals (Vancouver General Hospital and St. Paul's Hospital).
• Evaluated forecasting accuracy against actual ward-level capacity decisions made by hospital staff during the first three pandemic waves.

Main Results:

• Both statistical and ML forecasting methods demonstrated valuable ward-level forecasting capabilities for pandemic resource planning.
• Forecasting methods using point forecasts with upper 95% prediction intervals showed higher accuracy in anticipating required beds compared to hospital staff's decisions.
• The developed methodology was integrated into a publicly available online tool for operationalizing ward-level forecasting.

Conclusions:

• Ward-level forecasting tools utilizing statistical and ML methods can significantly enhance hospital resource planning during pandemics.
• The implemented tool can empower hospital staff to make more informed decisions, leading to improved patient care and reduced staff burnout.
• The publicly available tool offers a practical solution for optimizing hospital capacity planning in response to public health emergencies.