A large review of the results from rapid point-of-care COVID-19 tests found inconclusive evidence of their accuracy and led to a plea for more studies.

The researchers at Cochrane sought to determine if rapid point-of-care antigen and molecular tests for COVID-19 are accurate enough to replace RT-PCR tests, which require laboratory analysis to diagnose infection.

They reviewed 22 publications reporting on 18 study cohorts with 3,198 unique samples, of which 1,775 had confirmed SARS-CoV-2 infection. Ten of those studies took place in North America, with two in South America, four in Europe, one in China and one conducted internally.

The results came from eight commercial tests (four antigen and four molecular) and one in-house antigen test. They included considerable variation between the outcomes of the antigen tests and how well they detected COVID-19 infection. Antigen tests gave false-positives in less than 1% of samples. Evaluations of molecular tests found that they correctly detected an average of 95% of samples with COVID-19 infection, with around 1% yielding false-positives.

The researchers said that their confidence in the evidence is limited. Of the 18 relevant studies they analyzed, nine deliberately included a high percentage of people with confirmed COVID-19 infection or exclusively people with COVID-19. Information regarding the people providing the samples for testing was not available for 14 of the studies and 12 did not offer any information about where people were tested. None included samples from people without symptoms.

About three-quarters of studies did not follow the test manufacturers’ instructions, the researchers reported. Often, studies did not use the most reliable methods or report enough information, and a quarter of the studies were published early as “preprints” that did not undergo the normal rigorous checks of published studies, according to the review, which includes data up to May 25, 2020.

The researchers concluded that, because of the dearth of information regarding participants in the studies, the evidence is not strong enough and said more studies are “urgently needed” so that the tests can be evaluated for their effectiveness in practice.

With new research being published in the field, the researchers noted that they plan to update the review soon.