Identifying real time surveillance indicators to estimate COVID-19 hospital admissions in Colorado during and after the public health emergency

Abstract

Questions remain about how best to focus surveillance efforts for COVID-19 and other emerging respiratory diseases. We used an archive of COVID-19 data in Colorado from October 2020 to March 2024 to reconstruct seven real-time surveillance indicators. We assessed how well the indicators predicted 7-day average COVID-19 hospital admissions, a key indicator of outbreak severity, using machine learning and regression models, and used cross-correlation analysis to identify leading indicators. We found that hospital-based surveillance metrics, including real-time hospital census data and emergency-department based syndromic surveillance, were among the best predictors of COVID-19 hospital admissions during and after the public health emergency (PHE). While wastewater was a weaker individual predictor, its removal from our multi-indicator models resulted in a decrease in model performance, suggesting that wastewater provides important, unique information. Likewise, we found that test positivity, while imprecise, can serve as a leading indicator of COVID-19 hospitalizations. These findings suggest hospital-based reporting should be a surveillance priority, and that wastewater surveillance and test positivity can improve situational awareness for COVID-19 in Colorado. In contrast, case reporting was not found to be essential to real-time monitoring of COVID-19 hospitalizations in Colorado. The generalizability to other regions and respiratory illnesses warrants further investigation.