[ad_1]
One in five people who undergo rapid tests for Covid could be incorrectly told they have the virus, according to new research suggesting that accuracy rates may be lower than initially thought.
The study, published Thursday in the British Medical Journal (BMJ), examines the performance of fingerstick antibody tests for SARS-CoV-2, the virus responsible for Covid-19 infection.
The results suggest that if 10 percent of the people tested had already been infected, about 20 percent of all positive results would be incorrect or “false positives.”
The findings are significant in that they contradict those of previous, non-peer-reviewed research that found there were no inaccuracies among the positive tests.
Therefore, although they can offer a sufficient degree of precision for surveillance studies of a population, laboratory confirmation of positive results is likely to be required to provide evidence of any protection against the virus.
The research was conducted by scientists from Public Health England and the Universities of Bristol, Cambridge and Warwick.
They looked at the AbC-19 rapid test, which uses a drop of blood from a finger prick to assess the likelihood that the person has been previously infected, delivering results in 20 minutes without the need for laboratory tests. Such rapid change is invaluable in a world struggling to function in the ongoing pandemic.
In Ireland, there appears to be no prospect of any rapid test system being approved this year.
The investigation analyzed blood samples in a laboratory of 2,847 key health care, fire and police workers since last June.
Of these, 268 had a previous positive result and were classified as “known positives”. The remaining 2,579 had an unknown previous infection status. A further 1,995 pre-pandemic blood samples were also tested as “known negatives”.
They estimated that the ability of the AbC-19 test to correctly identify a true negative sample was 97.9 percent, meaning that 2.1 percent of people who had not had a prior SARS-Cov infection. 2 tested positive incorrectly.
They also estimated their ability to correctly identify a true positive sample at 92.5 percent based on confirmed cases, but considerably lower (84.7 percent) in those with unknown prior infection status.
This difference, the researchers explain, is likely due to the fact that the test is more sensitive when antibody levels are higher: those with a positive PCR result tend to have more severe disease, meaning they are likely to have produced more antibodies.
The lower figure of 84.7 percent is believed to be a more realistic level of estimate, meaning the test would miss 15.3 percent of people with a prior SARS-CoV-2 infection.
That means if 1 million people were tested, 10 percent of whom had previously been infected with SARS-CoV-2, there would be 18,900 false positive results. Overall, about one in five positive results would be wrong, they said.
The test’s accuracy could be even lower if used by members of the public at home, the research found.
In a linked editorial, Dipender Gill at Imperial College London and Mark Ponsford at Cardiff University noted that “a clear message must be communicated to the public that positive results from these trials do not provide evidence of immunity.”
