Accurately predict short-term Covid-19 demand for general hospital and critical care beds
CoviDemand uses locally reported admissions of Covid-19. Predictions are updated frequently based on daily reports from the Trust. Diagnostic uncertainty, such as where testing results are not concurrently available, is modelled probabilistically.
The model builds on local knowledge, including, for example, closed wards or flexible bed stock. This gets considered when assessing capacity bottlenecks.
Example problem solved:
• Demand for critical care beds from people that are critically ill with Covid-19
• Demand for high oxygen supply hospital beds from people that are seriously ill with Covid-19
• Demand for general hospital beds from people that are hospitalised with Covid-19
• Summary tables that are easy to interpret by operational teams
• Outputs are broken down by critical, high oxygen and general acute beds.
• Two-week ahead predictions
• Tried and tested successfully during March, April, and May
• Clear outputs that can be used by operational teams
• Support improved Covid-19 response by understanding the expected demand