Covid-19 Hospitalizations and Deaths Predicted by Sars-Cov-2 Levels in Boise, Idaho Wastewater

The viral load of COVID-19 in untreated wastewater from Idaho’s capital city Boise, ID (Ada County) has been used to predict changes in hospital admissions (statewide in Idaho) and deaths (Ada County) using distributed fixed lag modeling and artificial neural networks. The wastewater viral counts were used to determine the lag time between wastewater virus counts and COVID-19 hospitalizations as well as deaths (18 and 23 days, respectively). Quantitative measurement of SARSCoV-2 viral RNA counts in the untreated wastewater was determined using RT-qPCR over a span of 13 months. To mitigate the effects of PCR inhibitors, a series of dilution tests were conducted, and the 1/4 dilution generated the most successful model. To the best of our knowledge, this is the first report of prediction of deaths from Covid-19 based on wastewater RNA counts using machine learning-based multilayered artificial neural networks. The applied modeling demonstrates that environmental surveillance data can be combined with hospital admissions and death data to generate machine learning-based artificial neural network models that predict COVID-19 hospital admissions and deaths, providing an early warning for medical response teams or health department policymakers.