Using publicly available data from emerging research on COVID-19, this brief was written and reviewed by the Coronavirus-19 Outbreak Response Experts (CORE-19) at the University of Tennessee, Knoxville. It provides information on the accuracy of COVID-19 testing.
False Negative and False Positive Rates of COVID-19 Test
Since there is currently no vaccine or drug(s) for preventing or treating COVID-19, it is essential to detect the disease accurately and at an early stage, and immediately isolate the infected persons from the rest of the population.
To determine whether a suspected person has COVID-19 or not, samples taken from the back of a person’s nose or throat are collected and submitted to a lab for detection of genetic material from the coronavirus.
Any diagnostic test for COVID-19 must correctly identify those with the disease as having the disease (sensitivity of the test) and those without the disease as not having the disease (specificity of the test).
In the US, the most commonly used test is the reverse transcription polymerase chain reaction (RT-PCR). This, and a variety of other tests using a slightly different PCR platform were approved by the United States Food and Drug Administration (FDA) and rapidly deployed under the Emergency Use Authorization (EUA). Currently, there are no published data on the performance of these tests in real-world situations.
These data imply that 27-46% of the people whose nasal specimens were assessed, and 20-70% of those whose throat swabs were assessed for COVID-19 were inaccurately told that they did not have the disease when they actually had it.
Given the highly contagious nature of COVID-19 (on average, each infected person can infect 3 other people, the low sensitivity for the test underscores the need for continued social distancing and other measures to combat the spread of the disease in the community and ensure that hospitals and centralized quarantine facilities do not get overwhelmed.
The low sensitives also show the need for caution in interpretation test results; if a person receives a negative RT-PCR test but suspicion for COVID-19 remains, the person should be isolated and re-sampled several days later.
A number of studies suggest that chest CT may complement clinical symptoms and genetic tests in early diagnosis of COVID-19. For instance, Ai and his colleagues demonstrated higher sensitivity (97-98%) for chest CT, a routine imaging tool for rapid diagnosis of pneumonia, than for RT-PCR test (66-80% sensitivity for throat swabs).
Moreover, they observed lung abnormalities typical of COVID-19 in chest CTs of patients with COVID-19 clinical symptoms but negative RT-PCR results. Xie and his colleagues reported negative RT-PCR tests for throat swabs despite chest CT findings suggestive of viral pneumonia in some patients who ultimately tested positive for the COVID-19 virus.
Among persons with a positive test result for COVID-19, there is a chance that some of them may not actually have the disease i.e. they may be false positives. There is no published data regarding the false positive rates for the myriad of tests being used across the US.
How long, after exposure, does it take for the COVID-19 virus to be detectable?
At this time, no studies have been done to investigate how long, after infection, it takes for the virus to be detected. However, sampling early in the disease course may lead to a negative RT-PCR test compared to sampling later.
For instance, Ai and his colleagues reported that 23% of patients with suspected COVID-19 tested positive after an initial negative PCR test. Viral genetic materials have been reported to be higher soon after symptom onset (0-7 days) compared with later in the illness (See Yang et al. and Zou et al.).
A study performed in China by Yang and colleagues detected viral genetic material in the upper respiratory tract samples collected as early as 3 days after onset of illness.
A list of testing site locations in Tennessee counties can be found here.
Coronavirus-19 Outbreak Response Experts (CORE-19)