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

The model runs in discrete half-day time steps, representing day and night. During the day, people go to institutions (mixing groups) appropriate to their age, and at night they return to their families (where ages mix). Working-age adults may commute to different census tracts during the day. We do not distinguish between weekdays and weekends, so people go to school and work every day. Susceptible people become infected when they are in the same setting as infected people. Susceptibles also have a small chance of infection if they are in the same community at the same time as an infectious person. Children and adults are equally susceptible to infection, but children tend to have stronger "contacts" in the model so are more likely to infect and become infected than adults. Upon infection, a person may become infectious starting the next day (regardless of symptom status) and their level of infectiousness can change daily (e.g., exponential, log-normal, etc). A fraction of infectious people become symptomatic after an incubation period, which is specified as a CDF for flexible parameterization (e.g., can be fixed incubation time, normal, Weibull, etc). Symptomatic people may choose to stop going to work or school because of illness. Public health interventions, such as home quarantine or school closures, may be based on the cumulative prevalence of detected symptomatic cases. Social distancing is implemented by closing settings, like schools, to stop transmission but this is partially offset by these individuals going elsewhere. Much more detail about model structure and assumptions is in [13] .