Abstract
Since SARS-CoV-2 emerged, a 14-day quarantine has been recommended based on COVID-19”s incubation period. Using an RT-PCR or rapid antigen test to “test out” of quarantine is a frequently proposed strategy to shorten duration without increasing risk. We calculated the probability that infected individuals test negative for SARS-CoV-2 on a particular day post-infection and remain symptom free for some period of time. We estimate that an infected individual has a 20.1% chance (95% CI 9.8-32.6) of testing RT-PCR negative on day five post-infection and remaining asymptomatic until day seven. We also show that the added information a test provides decreases as we move further from the test date, hence a less sensitive test that returns rapid results is often preferable to a more sensitive test with a delay.
Background
Quarantine of potentially infected individuals is a time-tested approach for controlling epidemics. Since SARS-CoV-2 emerged, the US Centers for Disease Control and Prevention (CDC), World Health Organization and others have recommended a 14-day quarantine based on COVID-19”s incubation period.
Concerns have been raised that this is overly burdensome and provides limited added benefit versus shorter durations that may increase compliance by reducing disruptions to normal activities (e.g., work, school). Using an RT-PCR or rapid antigen test to “test out” of quarantine is a frequently proposed strategy to shorten duration without increasing risk. The CDC recently issued guidance that a 7-day quarantine with a negative RT-PCR test from a sample collected within 48 hours of exit, or a 10-day quarantine with no test, may sometimes be acceptable.(1) However, RT-PCR tests can miss infections, particularly prior to symptom onset.(2) The goal of our study was to quantify the proportion of infections that escape detection under different testing strategies and varying quarantine durations.
Methods
We calculated the probability that infected individuals test negative for SARS-CoV-2 on a particular day post-infection and remain symptom free for some period of time. This is equivalent to the probability that an infected individual leaves quarantine given a particular test time and quarantine duration. See supplement for details.
To calculate these probabilities, we developed a biologically plausible model assuming that test sensitivity and symptom onset are independently linked to an unobserved time point after infection where viral load crosses a critical threshold (Figure 1). Available sensitivity data is usually indexed to time of symptom onset or an inferred time of infection.(2–4) Using such data directly tacitly assumes test sensitivity is either fully dependent on, or fully independent of, symptom onset timing. More likely, both depend on infection timing and an underlying process of viral replication, and thus, as our model assumes, are partially correlated. We obtain estimates of sensitivity conditional on the timing of infection and symptom onset using empirically derived distributions of the incubation period(5) and test performance(2).
Results
An RT-PCR test after the first few days of quarantine appreciably decreases the likelihood of an infected individual going undetected (Figure 2A). However, the added information decreases as we move further from the test date, hence a less sensitive test that returns rapid results is often preferable to a more sensitive test with a delay (e.g., Figure 2B). We estimate that an infected individual has a 20.1% chance (95% CI 9.8-32.6) of testing RT-PCR negative on day five post-infection and remaining asymptomatic until day seven, similar to an 8-day quarantine without a test. A negative test on day seven and remaining asymptomatic until day eight misses 9.4% of infections (95% CI 3.1-20.1), similar to a 10-day quarantine, and a negative test and no symptoms on day 10 misses 2.2% (95% CI 0.5-7.4), similar to a 14-day quarantine.
Discussion
A negative test can substantially increase the confidence that an individual in quarantine is uninfected, potentially achieving an impact similar to that of quarantines 3-4 days longer. However, the importance of a negative test decreases each day an individual goes without developing symptoms after the test. Hence, a less sensitive test that returns rapid results, and can be taken closer to the end of quarantine, may be more beneficial than a more sensitive test with a delay. The importance of timely test results has been shown to be critical in other contexts(6), and is one reason a negative test on day 5 adds little benefit to 7 days of quarantine without a test.
Data Availability
Data and R code needed to reproduce the results are provided.
https://lucymcgowan.github.io/covidsens/articles/articles/figures.html
Footnotes
This work was supported by funding from State of California Institute of Technology (19-13081) (ECL, KHG, JL), Johns Hopkins Hospital (ECL, JL), and Bloomberg Philanthropies (ECL, KHG, ESG, JL)