Real-time forecasting of COVID-19 bed occupancy in wards and Intensive Care Units

Health Care Manag Sci. 2021 Jun;24(2):402-419. doi: 10.1007/s10729-021-09553-5. Epub 2021 Mar 25.

Abstract

This paper presents a mathematical model that provides a real-time forecast of the number of COVID-19 patients admitted to the ward and the Intensive Care Unit (ICU) of a hospital based on the predicted inflow of patients, their Length of Stay (LoS) in both the ward and the ICU as well as transfer of patients between the ward and the ICU. The data required for this forecast is obtained directly from the hospital's data warehouse. The resulting algorithm is tested on data from the first COVID-19 peak in the Netherlands, showing that the forecast is very accurate. The forecast may be visualised in real-time in the hospital's control centre and is used in several Dutch hospitals during the second COVID-19 peak.

Keywords: Bed occupancy; COVID-19; Forecast; Kaplan-Meier estimator; Network of infinite server queues; Richards’ curve.

MeSH terms

  • Bed Occupancy / trends*
  • COVID-19*
  • Forecasting
  • Hospitals
  • Humans
  • Intensive Care Units*
  • Kaplan-Meier Estimate
  • Models, Statistical
  • Netherlands
  • SARS-CoV-2