Application of a Serial Antigen Based Testing Strategy for SARS-CoV-2 and Student Adherence in a University Setting, Wisconsin — October–November 2020

Background Serial severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) testing has been implemented at institutions of higher education (IHEs) and other settings. Testing strategies can include algorithms specifying confirmatory reverse-transcription polymerase chain reaction (RT-PCR) testing after an antigen test. It is unknown how testing strategies perform detecting SARS-CoV-2, including individual adherence to serial testing requirements. Methods Student serial testing adherence was defined as completing ≥80% of weekly tests from October 5, 2020 to November 14, 2020 and evaluated using logistic regression. Medical records were reviewed for all positive antigen test encounters and 10% of daily negative antigen test encounters during October 19-November 30, 2020. Results were used to estimate the proportion of individuals requiring only antigen tests, requiring and completing RT-PCR testing, and associated costs of tests. Results Two thirds (66.5%; 1166 of 1754) of eligible on-campus students adhered to weekly testing; female students were more adherent (adjusted odds ratio [aOR], 2.07; 95% confidence interval, 1.66-2.59) than male students. Of all antigen test encounters, 11.5% (1409 of 12 305) reported >1 COVID-19 symptoms. Of non-COVID-19-exposed antigen test encounters, 88% (10 386 of 11 769) did not require confirmatory RT-PCR testing. Only 28% (390 of 1387) of testing encounters had an associated recommended confirmatory RT-PCR test performed. We estimated the testing strategy captured 61% (235 of 389) of predicted RT-PCR-positive specimens. Conclusions At this IHE, most students voluntarily adhered to serial testing. The majority of antigen test results did not require confirmatory RT-PCR testing, but when required, most students did not obtain it. Including strategies to increase the proportion of individuals obtaining indicated confirmatory testing might improve the testing program's performance.