As the rate of COVID-19 testing slowly creeps up in the US, public health experts have a new concern that many people with negative test results actually have the virus.
“If you have had likely exposures and symptoms suggest Covid-19 infection, you probably have it — even if your test is negative,” wrote Harlan Krumholz, a professor of medicine at Yale, in The New York Times.
Doctors and clinicians struggle with test accuracy all the time, across all areas of medicine. No test is perfect. Under normal circumstances, though, they understand the factors that contribute to false negative or false positive results from a particular test. They might also have more extensive data on the test that helps them interpret what it says. Not in this case.
Why do false negatives exist?
“There’s a lot of talk saying ‘it’s a bad test’. I think it’s not that the test is bad,” says Catherine Klapperich, director of the Laboratory for Diagnostics and Global Healthcare Technologies at Boston University. Instead, she says, the health care providers and patients don’t have the information they need to fully understand their test results.
The bulk of the tests done in the US for COVID-19 use a technique called PCR, which looks for bits of the new coronavirus in a mucus sample taken from a patient. PCR works well, and it will flag a sample as positive even if there are only a few copies of the virus in it.
The problem is that the virus doesn’t tend to stay in an easily accessible part of the body. It lurks in the nasopharynx, where the back of the nose meets the top of the throat. To test someone for the new coronavirus, doctors and nurses have to stick a very long swab very deep into their nose. “It’s not rocket science, but you have to be trained to do it,” Klapperich says. Many false negative test results are probably because the swab wasn’t done correctly.
Doctors also don’t know when in the course of a COVID-19 illness the test works best. The data on the test’s false negative rate jumbles together all of the tests that have been done. It hasn’t broken out the false negative rates of tests done at different times during the progression of the disease. The false negative rate for tests done right when someone starts feeling sick, for instance, might be different than that same rate for tests done when people are hospitalized.
False negatives are not created equal
All tests are wrong sometimes, but clinicians are more comfortable with false results for certain types of tests than they are others. “There are variables that affect your tolerance for false negatives and positives,” Klapperich says.
On a screening test for HPV, a virus that can lead to cervical cancer, a false positive result is usually less dangerous than a false negative result. Someone with a positive test result will have additional follow-up tests to confirm if they actually have HPV and if they need additional treatment. In that case, if the positive result is incorrect, that can be corrected. If someone tests negative incorrectly, it could delay treatment. The anxiety and unnecessary follow-up tests that can come from a false positive result can cause harm. But for HPV, it’s not as risky as a false negative.
With the new coronavirus, it’s the opposite. If someone is told they have COVID-19, they’ll be told to quarantine. “They’ll be alone, and stressed — but safe. If you tell them they’re negative and they’re not, they could infect other people,” Klapperich says. As the consequences of this pandemic keep changing, health care workers treating patients with COIVD-19 are constantly reevaluating their tolerances for false positive or false negative testing results, she says.
Doctors have to decide if they can trust a negative test enough to stop wearing protective equipment when treating a hospitalized patient or if the clinical symptoms look enough like COVID-19 that the negative result doesn’t matter.
Normally, retesting sick patients could be a straightforward way to compensate for a less-accurate test. For something like a strep test, when a result doesn’t match a patient’s symptoms, a doctor can do a second type of test or a repeat test. Limited testing resources, though, make that much more challenging for COVID-19.
“Ideally, if someone tests positive, you’ll say they’re positive. If they’re negative and have symptoms, they could get another test. We can’t do that now,” Klapperich says. “We don’t have luxury of rerunning a test or sending someone for a test that’s complementary to get more data.”
Instead, doctors and patients have to decide on the fly what to do with a single negative or positive COVID-19 test. When they have more experience with a test, they’re better equipped to make those decisions. “Mammograms are good examples,” Klapperich says. False positives on those tests, which screen for breast cancer, happen fairly often. “People have the experience to say, ‘oh, you have a spot.’ Clinicians are trained to say that this is usually not a big deal.”
What comes next?
The coronavirus test is much newer than mammograms or tests for strep and HPV, and clinicians don’t have as much clear data to inform their interpretation of results. “There’s an interplay between the test and how well it does, and how people receive the test results,” Klapperich says. “Do they trust them? Do they trust the guidelines that go along with the test?”
Klapperich thinks there will be better information for both patients and providers available soon. Many clinicians are keeping good records and storing patient samples after they’ve been tested — while carefully noting when in the course of an illness the sample was taken. Soon, she says, they’ll be able to figure out how accurate the test is at different points in a case of COVID-19. That should help doctors make more confident recommendations that incorporate both when a test was done and what a patient’s symptoms are.
“The limitation right now is that people doing the testing are focused on patients,” she says. “When things settle down, and they don’t have to focus on patients every minute, they’ll do those studies.”