Coronavirus: Trump celebrates ‘testing milestone’ as CDC says antibody tests can be wrong half the time

While more than 15 million tests have been carried out in the US, the overall number hides myriad ambiguities

In an early-morning tweet on Wednesday, Mr Trump once again ballyhooed the US’s supposedly world-beating coronavirus testing achievements: “We pass 15,000,000 Tests Today, by far the most in the World. Open Safely!” He did not mention that the US is soon to reach 100,000 confirmed Covid-19 fatalities — or that testing data may come with any qualifications.

However, the US Centers for Disease Control (CDC) has issued new guidance emphasising that because data on reinfection is so far lacking, a positive test for antibodies should not be used to determine whether an individual is immune to the virus — or whether it is safe to lift social distancing requirements affecting any given population.

“Antibodies most commonly become detectable 1-3 weeks after symptom onset,” reads the guidance, “at which time evidence suggests that infectiousness likely is greatly decreased and that some degree of immunity from future infection has developed.

“However, additional data are needed before modifying public health recommendations based on serologic test results, including decisions on discontinuing physical distancing and using personal protective equipment.”

Most candidate coronavirus vaccines are only just entering human trials and pressure to lift local and national lockdowns is only increasing, meaning that accurate testing on a large scale is a key part of monitoring the spread of the virus as everyday interaction resumes. The implications of the guidance are potentially serious.

In particular, antibody testing is needed to determine what percentage of the population have already been exposed to the virus — regardless of whether they have shown symptoms — on the assumption that those who have been infected are now immune.

But, according to the CDC guidance, the rate of exposure to the virus in most places is expected to be below 25 per cent of the population. And depending on just how low that figure is, the accuracy of the tests can suffer greatly: “In a population where the prevalence is 5%, a test with 90% sensitivity and 95% specificity will yield a positive predictive value of 49%.

“In other words, less than half of those testing positive will truly have antibodies.”

There are multiple other complexities at work, including that there are many different tests available for use and that various states and agencies have been reporting their testing data in confusing ways, including by combining antibody test results with viral test results — making it hard to plot how the number of current cases is rising or declining over time.