Defining alarm thresholds for the load of pathogenic viruses in wastewater for decision making: An application to...

Wastewater monitoring has the potential to complement infectious disease surveillance systems. However, the absence of predefined viral signal thresholds in wastewater is often presented as a limiting factor in triggering public health action. To overcome this issue, the feasibility of defining alarm threshold for viral loads in wastewater samples was assessed by quantifying genome fragments of SARS-CoV-2, influenza A virus (IAV), respiratory syncytial virus (RSV), norovirus (NoV), and rotavirus (RoV) by RT-digital PCR (dPCR) in untreated wastewater samples from three treatment plants. Cut-point values were calculated for periods with a high rate of visits to emergency rooms or at-home visits by SOS Médecins for the related diseases. ROC curves were constructed, and the values of alarm threshold in wastewater were defined using the Youden index. For each targeted virus, alarm thresholds were close to each other across the three WWTPs. As indicated by likelihood ratios, evidence to rule in the diagnosis of high rate of visits when the alarm threshold was exceeded ranged from weak to strong and was highest for RSV and SARS-CoV-2. Evidence to rule out the diagnosis when the alarm threshold was not exceeded was strong or moderate for IAV, SARS-CoV-2 and RSV. Diagnostic performance of the test was not as high for NoV and RoV. Positive predictive value was highest for SARS-CoV-2 and RSV. For SARS-CoV-2 and RSV, the definition of an alarm threshold in wastewater could substantially inform the diagnosis of a period with a high rate of medical visits for COVID-19 and bronchiolitis, respectively.