Modeling layered non-pharmaceutical interventions against SARS-CoV-2 in the United States with Corvid

Dennis L Chao; Assaf P Oron; Devabhaktuni Srikrishna; Michael Famulare

Testing and home isolation of positive cases could serve the broader community more than the families of these cases. However, our model highlights that untested family members may often be infected before index cases are tested, so automatic home quarantine for all family members might be warranted if testing them is not possible. Because most infections in our model come from pre-symptomatic people or from people who never show symptoms, it may be impossible to prevent most infections with test-and-isolation policies unless non-symptomatic people are also isolated (e.g., quarantine of families or testing non-symptomatic people). The testing and isolation simulation results are particularly sensitive to assumptions about asymptomatic and pre-symptomatic transmission [16] , neither of which is yet well-characterized for SARS-CoV-2. Another modeling study found testing with isolation to be more effective, but their model assumed less asymptomatic transmission than Corvid does currently [17] . Contact tracing might be required to control SARS-CoV-2 outbreaks [18] , but could be more labor-intensive than implementing quarantine of families of ascertained cases. The proportion of SARS-CoV-2 infections that become symptomatic appears to be age-dependent [19] [20] [21] , but this is both hard to measure and hard to incorporate in a model. In Corvid, the presence of symptoms drives testing of cases, and mildly symptomatic people might not be suspected of having COVID-19. We therefore need to carefully consider how to choose an appropriate symptomatic fraction to model the human behaviors that drive both disease spread and detection. More extensive testing of asymptomatic contacts would help us understand the prevalence of asymptomatic infections and the spectrum of disease more generally.