“Every scientist and his dog is trying to make this test,” said James Gill, a clinical lecturer at Warwick Medical School in Coventry, England. “Whichever one is the fastest, whichever one is the cheapest, and whichever one is most reliable will be adopted.”
PharmAct’s test, intended for healthcare professionals, suffered a public relations blow shortly after it went on sale last month, when a prominent virologist told a German newspaper that he had used it for research in the country’s biggest coronavirus hotspot and that it had failed to catch two-thirds of cases.
“That hurt us a lot,” Burgard said. “It was written all over that these tests are rubbish.” He calls that analysis “bad science,” arguing that his company’s test can’t identify infection early on, since people have yet to build up antibodies. The virologist’s sample size was also too small to be instructive, he said.
[...]
Most test developers tout their products as performing very accurately when properly used, according to Marco Donolato, co-founder and chief scientific officer of BluSense Diagnostics ApS in Copenhagen. “If you see the brochure of every company, it’s 99%,” Donolato said.
In reality, the tests often don’t perform so well in the hands of inexperienced users who are often toiling in sub-optimal conditions. BluSense is working to validate its own antibody test with blood samples in Denmark. The process is often iterative, with test-makers tweaking their kits gradually to improve accuracy.
An impartial review of dozens of antibody tests will be available in coming weeks from the Foundation for Innovative New Diagnostics, a nonprofit group based in Geneva. PharmAct’s test isn’t among them, though the organization plans to keep investigating more candidates on a rolling basis.
From the article:
[...]