To compare the effectiveness and cost-effectiveness of SARS-CoV-2 screening strategies
using rapid antigen tests in a residential college campus.
A compartmental epidemic model was developed from published sources, primarily from
Paltiel et al. JAMA Network Open 2020. In model development we identified several
needed corrections to the published model which were confirmed with the lead published
author. Our model was verified against all published numeric outcomes of Paltiel (2020).
Our modeling study included a hypothetical cohort of 5000 students – 4990 students
without SARS-CoV-2 infection and ten with undetected, asymptomatic SARS-CoV-2 infection.
The screening strategies that we evaluated were symptom-based screening and tests
of varying frequency (i.e., every 1, 2, 3, and 7 days) and sensitivity (i.e., 40,
50, and 60%). These three levels of sensitivity were based on published values for
rapid antigen tests, e.g., 35.8% for asymptomatic persons and 64.2% for symptomatic
persons (Prince-Guerra et al. MMWR Morb Mortal Wkly Rep. 2021). Specificity was set
to 98%, the test cost to $10, and reproductive number (Rt) to 2.5. Model projections
were for an 80-day, abbreviated semester.
Within each sensitivity level (i.e., 40, 50, and 60%), screening frequency order of
every 7, 3, 2, and 1 days was associated with increasing costs (range $682,000-$4,639,200)
and increasing infections averted (range 1202-4797). In terms of cost-effectiveness,
for sensitivity levels 40 and 50%, the 7-day screening intervals were dominated. For
each sensitivity level, the incremental cost-effectiveness ratio (ICER, $ per infection
averted) increased with increased testing frequency. In general, as the test sensitivity
increased, ICERs increased, i.e., at higher sensitivities the cost to avert an infection
increases as the test frequency increases.
From an effectiveness standpoint, more frequent testing is preferable. From a cost-effectiveness
standpoint, even for poorly sensitive tests, screening every three days is an attractive