Can you trust COVID-19 test results?

False positives unlikely with PCR tests, experts say

Every time we journalists write about new COVID-19 cases and hospitalizations in Colorado, some variation of the same objection comes in: we’re making too much of nothing.

At first, it was that COVID-19 is no worse than flu. That’s been thoroughly debunked, as the new coronavirus caused more hospitalizations and deaths than any flu season in recent memory — though for unknown reasons it does seem to be less risky than the flu for toddlers.

Then it was that so many infections cause mild or no symptoms that reporting on increases in cases was pointless. But while the number of reported cases can be influenced by how much testing a state does, each new wave of increased cases in Colorado has been followed by increased hospitalizations and deaths.

Now, it’s claims that testing is so unreliable that it’s not worth talking about increases in cases. But that’s not true either.

Colorado only reports polymerase chain reaction (PCR) tests, which look for genetic material specific to the virus that causes COVID-19. Basically, through a series of chemical reactions, the test amplifies any genetic material from the virus until there’s enough to detect a signal. PCR tests are the most reliable type.

If you let the reactions run indefinitely, you could start generating false positives, but every brand of test comes with a cut-off point and guidelines to interpret the results, said Dr. David Kroll, a professor of pharmaceutical science the University of Colorado Anschutz campus. The testing kits also come with positive and negative samples, to make sure the process is working correctly before testing someone’s nasal swabs, he said.

“These are exhaustive tests,” he said.

Some PCR tests could pick up “broken pieces” of genetic material when a person is no longer contagious, Kroll said, though others look for more genetic markers and are less likely to rate those samples as positive. That’s why the Centers for Disease Control and Prevention advises most people to isolate for 10 days after their symptoms clear up, rather than getting retested repeatedly, he said.

It’s far more likely to get a false negative than to test positive when you’re no longer contagious, though, Kroll said. If you get tested too soon after exposure, or the person collecting your sample doesn’t take enough from your nostrils, the test could miss the virus because there aren’t enough particles to show up as positive by the cut-off point, he said.

There isn’t enough research yet to be sure what level of virus in a sample means someone is no longer contagious, said Jim Dunn, director of medical microbiology and virology at Texas Children’s Hospital. If you feel fine but test positive, you have no way of knowing if you’re in the early stages of the disease, so the only way to ensure you don’t infect anyone is to isolate yourself until the contagious period passes, he said.

“I think it’s better to err on the side of caution,” he said.