Antibody tests used to confirm whether someone has ever had the coronavirus could be wrong half of the time, driving bad policy decisions, the US Centers for Disease Control and Prevention warns.
The Centers for Disease Control and Prevention (CDC), a top federal health agency, said in a new guidance on Tuesday that during mass screenings in various populations, up to half of the positive results from antibody tests - also known as serologic tests - could be false, warning there are still major obstacles in determining who has been infected or what immunity they might have.
The unreliable tests shouldn't inform decisions about moving potential patients to "congregate settings" - such as schools, dormitories or jails - nor choices "about returning persons to the workplace," the CDC said.
But even when a patient is "truly positive" for antibodies, the agency said the result still doesn't prove that the individual is "protected from future infection" or immune to the virus.
Serologic testing should not be used to determine immune status in individuals until the presence, durability, and duration of immunity is established.
The rate of false positive and negative test results heavily depends on the overall number of infections in a population. If a relatively low proportion of people in a given area have contracted the virus - five percent, for example - the CDC says even a highly sensitive and specific test will still yield a significant number of false positives, up to 51 percent.
In other words, less than half of those testing positive will truly have antibodies.
Echoing the new guidance, a research team at the University of Minnesota has also warned about antibody tests, issuing a report last week cautioning that they should not be used as a "diagnostic tool" given their inaccuracy.
To get around some of the shortcomings, the CDC recommended that health professionals use only the highest quality tests with the greatest level of specificity, conduct repeat screenings of those who come up positive to verify their result, and to focus testing on patients with a history of Covid-19-like symptoms.
The CDC announcement comes less than a week after the agency admitted that it has been combining data from antibody and RNA tests, conflating two incompatible metrics - one that tallies active infections, and another that looks to see whether someone has ever had the virus. The blending of data, first reported by the Atlantic, could seriously skew figures used by state governments to chart a course to reopening. And given the CDC's latest announcement about antibody tests, the overall numbers may have been distorted even further by false positives.
In addition to the CDC itself, at least four state-level health agencies - in Texas, Georgia, Pennsylvania, and Vermont - said they are combining figures in the same way, according to the Atlantic.
The health agency also recently made waves when it revised its overall case fatality rate (CFR) estimate, putting the number around 0.4 percent, nearly 10 times lower than the World Health Organization's initial figure of 3.4.
Despite the questionable data, however, dozens of US states are now moving to lift sweeping containment measures imposed in March to stem the spread of the virus, as some of the nation's worst hot spots appear to be subsiding, most notably New York City. To date, nearly 1.7 million people have tested positive for the coronavirus in the US throughout the pandemic, with more than 98,000 fatal cases, according to data collected by Johns Hopkins University.
Think your friends would be interested ? Share this story !