COVID-19 Epidemiological Projections for Penn State’s University Park Campus Under its Fall 2021 Plan and Some Alternatives

We present new COVID-19 epidemiological predictions for PSU University Park campus for the Fall 2021 semester. The basic model is a Susceptible-Exposed-Infected-Recovered (SEIR) compartmental model, updated to account for vaccinations. The original version of the model was previously used to model the Fall 2020 semester. We use the updated model to evaluate alternative scenarios, within the testing capacity at UP, comparing with the current PSU administration’s COVID-19 plan. We highlight three key conclusions:

• It is important to mandate vaccinations so as to reduce symptomatic cases and possible hospitalizations and deaths among students, faculty, and staff;
• It is urgent to start testing all the vaccinated individuals, a strategy totally absent in the present PSU plan, and
• It is necessary to implement mitigation measures during home football games, such as mandatory submission of negative COVID-19 test results for all visitors to Beaver stadium, or, if this is not feasible, to hold games with a much reduced attendance.

We have carried out Monte Carlo simulations of the epidemiological model that account for the uncertainty in some key model parameters. As expected, an overall percentage of vaccination in the mid 90%s, achievable only with a vaccine mandate, notably reduces the number of moderately symptomatic infections and therefore the possibility of hospitalizations and deaths among students, staff and faculty. Without also testing the vaccinated, however, the current PSU plan will not be able to stop an outbreak of asymptomatic infection at UP even if a vaccine mandate is rapidly implemented. Under the current PSU plan, an outbreak of asymptomatic infections, driven by infections among vaccinated individuals is likely to occur and will likely go totally unnoticed. The infections may then spread into the surrounding community and affect the unvaccinated population in Centre County, most notably children under the age of 12. We analyze the test results data available after 12 days in the semester in the PSU dashboard and contrast it with the model predictions.