COVID-19 screening strategies that permit the safe re-opening of college campuses

Abstract

Importance: The COVID-19 pandemic poses an existential threat to many US residential colleges: either they open their doors to students in September or they risk serious financial consequences.

Objective: To define SARS-CoV-2 screening performance standards that would permit the safe return of students to campus for the Fall 2020 semester.

Design: Decision and cost-effectiveness analysis linked to a compartmental epidemic model to evaluate campus screening using tests of varying frequency (daily-weekly), sensitivity (70%-99%), specificity (98%-99.7%), and cost ($10-$50/test). Reproductive numbers Rt = {1.5, 2.5, 3.5} defined three epidemic scenarios, with additional infections imported via exogenous shocks. We generally adhered to US government guidance for parameterization data.

Participants: A hypothetical cohort of 5000 college-age, uninfected students. Main Outcome(s) and Measure(s): Cumulative tests, infections, and costs; daily isolation dormitory census; incremental cost-effectiveness; and budget impact. All measured over an 80-day, abbreviated semester.

Results: With Rt = 2.5, daily screening with a 70% sensitive, 98% specific test produces 85 cumulative student infections and isolation dormitory daily census averaging 108 (88% false positives). Screening every 2 (7) days nets 135 (3662) cumulative infections and daily isolation census 66 (252) with 73% (4%) false positives. Across all scenarios, test frequency exerts more influence on outcomes than test sensitivity. Cost-effectiveness analysis selects screening every {2, 1, 7} days with a 70% sensitive test as the preferred strategy for Rt = {2.5, 3.5, 1.5}, implying a screening cost of {$470, $920, $120} per student per semester. Conclusions & Relevance: Rapid, inexpensive and frequently conducted screening (even if only 70% sensitive) would be cost-effective and produce a modest number of COVID-19 infections. While the optimal screening frequency hinges on the success of behavioral interventions to reduce the base severity of transmission (Rt), this could permit the safe return of student to campus.

Figure 1.. Cumulative infections as a function of test sensitivity and frequency.

Over an 80-day horizon, for the (a) base case (R_t 2.5), (b) worst case (R_t 3.5), and (c) best case (R_t 1.5), these figures report cumulative infections (vertical axis; logarithmic scale) for tests with sensitivity ranging from 70-99% (horizontal axis). The colored lines denote different screening test frequencies (blue: daily screens; orange: every 2 day screen; gray: every 3 day screen; yellow: weekly screen).

Figure 2:. Projecting the required size of the isolation dormitory.

An isolation dormitory needs to be large enough to house students with false positive results (shaded gray), students with symptoms (shaded light blue), and students without symptoms who have received true positive results (shaded dark blue). Over the 80-day horizon (time on the horizontal axis), this figure depicts the number of students in the isolation dormitory (vertical axis, note the scales are different) by indication, using a 70% sensitive, 98% specific test, under the base case scenario (Rt = 2.5). The panels show results of screening at different frequencies: (a) daily screening; (b) screening every 2 days; (c) screening every 3 days; and (d) weekly screening. In Panels a through c, the effect of exogenous shocks (5 per week) is visible in the scalloped borders; this is less evident with weekly testing where the number of true positive cases masks the comparatively small impact of exogenous shocks.