The GW COVID-19 case management system embedded in the Student Health medical record system was used to evaluate the data we collected relevant to RAT testing for people who were in isolation. There were 955 cases identified between February 21 and April 14, 2022. Cases with the following attributes were removed from the analysis: a positive test uploaded more than 7 days into the isolation period (n=34), false positives (n=2), no initial positive listed in the patient chart (n=2), individuals lost to follow-up (n=8), individuals who uploaded a PCR for purposes of isolation clearance (n=3), those with a positive test within 90 days of a previous positive test (n=4), and those who were medically released before 10 days based on verbal confirmation of negative antigen tests (n=22). The final analytic sample was 887 RAT results among 880 cases.

Descriptive statistics were computed for all variables. Categorical variables, including campus residential living status and student or employee designation, were described using frequencies and percentages. Continuous variables, including age and number of days in isolation, were described using medians and IQRs. Next, we calculated positive percentages for the day a RAT was submitted to identify if individuals were eligible to be released before the end of the 10-day isolation period. Then, simple logistic regressions were run to compute the odds of uploading a RAT by individual independent variables, including campus residential living (ie, on or off campus), student or employee designation, age, and days in isolation. A final multivariable logistic regression model computed the odds of uploading a RAT with the independent variables simultaneously, including campus residential living status (on or off campus), student or employee designation, age, and days in isolation. SAS (version 9.4) was used for all analyses.

All community members provided informed consent to participate in the GW COVID-19 surveillance program, and the GW Institutional Review Board concluded that these were nonresearch-related activities.

This analysis sought to evaluate the utility of using RATs to end isolation periods more safely prior to the full 10 days of isolation. Rapid antigen tests are beneficial, as they provide valuable insight to guide isolation and treatment, allowing health care providers to release individuals from isolation when they are no longer contagious and maintain isolation for people who may still be infectious. In this analysis, the percentage of positive RATs uploaded on day 5 of isolation was 45.6% (118/259). If isolation were set to only 5 days, most likely a proportion of the 45.6% who uploaded a positive RAT would still be infectious and able to spread COVID-19 on campus and in the community despite advice to employ one-way masking [15]. This result is consistent with prior studies conducted in various settings ranging from Mayo Clinic Florida’s COVID Virtual Clinic to other universities in the United States, which have reported RAT positivity rates of 26%-68% at day 5 [1,6,10]. Of note, the percentage of RAT uploads that were positive at day 7 (99/210, 47.1%) is no lower than those at day 5, as opposed to day 10 and beyond, when only 11.1% (7/63) were positive. Even in the context of fairly high positive percentages, the RAT testing strategy resulted in 477 cases being cleared prior to day 10 of their isolation or 1547 days of lost productivity saved.

After adjusting for confounders, living status, days in isolation, and primary affiliation were all significantly associated with uploading a RAT. All those living on campus were students, and they may have had higher odds of uploading a RAT because they were provided RATs in person. Off-campus cases had the option to either pick RATs up from an on-campus location or have a testing kit mailed to them, possibly resulting in delays, which may have made it more challenging for certain off-campus cases to complete a RAT when recommended. Employees were more likely to upload RATs, possibly due to reasons such as wanting to return to work or running out of pandemic leave or sick time. Additionally, occupational health followed up with cases through phone calls multiple times during their isolation, whereas the student health center was unable to do so. Certainly, a testing out of isolation strategy should ensure that all members of the population have equitable access to RATs as well as follow-up medical supervision.

Although this is not the first study to analyze a COVID-19 RAT protocol, this study has two unique strengths. First, the sample size was significantly larger than several similar studies, which ranged from 40 to 323 [6,10,16,17]. The CDC published a study analyzing 3502 COVID-19 infections, with 729 antigen tests in their study [18]. Second, previous studies among institutions of higher education have focused solely on students [10], while this study was able to assess the experience with both students and employees. Additionally, the use of RATs allows for individuals to return to work or class sooner than they would otherwise be allowed to return. This benefit is particularly important for a university setting because missing a class can be detrimental to academic achievement and can increase stress during an already potentially stressful time [19].

There are several limitations to this study. First, this study relied on cases self-reporting the results of their RATs, and we do not know why 24% (211/880) of our cases did not do so (whether related to lack of access to the tests, opting not to do the tests, failure to follow instructions to upload the RAT test results, or having symptoms throughout, such that the test results would have been irrelevant). Almost certainly, more cases took a RAT than GW was aware of, which could skew the results, depending on why the RATs were not uploaded. For those with asymptomatic infections, either at the outset or over the entire course of the illness, it is difficult to ascertain when they acquired their infection [18]. Thus, asymptomatic cases may be more likely to have a negative RAT as additional days may have elapsed since the onset of their infection. Compared to prior studies [6,10,16,18], this study did not perform analyses based on vaccine status, as only 2 cases had vaccine exemptions.

The most important issue is how to best isolate people with COVID-19 to balance public health protection with sparing individuals’ needless days in isolation. This study, and many others, support the notion that automatically ending isolation at 5 days is risky in terms of sending many who are infectious back into the community, in fact, that 7 days may be too short as well. Yet many can come out of isolation earlier with RAT testing. Many questions remain, including the efficacy of one-way masking for those coming out of isolation who still are RAT positive (as occurs with the current CDC protocol) and the need to document transmission of COVID-19 from people who have been isolated fewer than 10 days. Additionally, for both asymptomatic and symptomatic cases, RATs may not be the best measure of infectiousness [6,15,20,21], possibly resulting in the premature release of cases from isolation or an unnecessary extension of isolation. These research avenues, along with findings from this study, can guide the development of better-informed isolation policies that are protective of public health and cause the least disruption to individuals’ lives.