Derivation and validation of a point-based forecasting tool for SARS-CoV-2 critical care occupancy: a...

Area of Science:

• Epidemiology
• Public Health
• Health Services Research

Background:

• Critical care resources were strained during the COVID-19 pandemic due to SARS-CoV-2 infections.
• A previous regression model forecasted critical care occupancy using case numbers, age, and testing volume.
• This study aimed to validate and update the forecasting model considering population immunity and vaccination.

Purpose of the Study:

• To validate and update a regression-based model for forecasting critical care occupancy.
• To incorporate the impact of population immunity and widespread vaccination into the model.
• To estimate vaccine-attributable reductions in critical care admissions for SARS-CoV-2.

Main Methods:

• Utilized provincial SARS-CoV-2 case, testing, and vaccination data from March 2020 to September 2022.
• Developed an initial model using data from the first two pandemic waves and an updated model including wave 3.
• Validated models by comparing projections to unused data and assessed predictive validity using Spearman's rho.

Main Results:

• The initial model showed good calibration but modest predictive validity (rho = 0.46).
• Predictive validity improved significantly with models incorporating wave 3 data and vaccination (rho = 0.68).
• Vaccination was associated with an estimated 144% reduction in expected critical care admissions.

Conclusions:

• Simple regression models are valuable for predicting SARS-CoV-2 critical care occupancy.
• Forecasting models developed early in the pandemic need recalibration to account for evolving immunity and vaccination.
• Updated models provide more accurate predictions and quantify the impact of vaccination on critical care demand.