Jacqueline Chu, M.D., considered the man with a negative coronavirus test on the other end of the phone, and knew, her heart dropping, that the test result was not enough to clear him for work.
The man was a grocery store clerk—an essential worker—and the sole earner for his family. A 14-day isolation period would put him at risk of getting fired or not having enough money to make rent that month. But he had just developed classic COVID-19 symptoms, and many others around him in Chelsea, Massachusetts, had confirmed cases. Even with the negative test, his chances of having the disease were too high to dismiss.
For many Americans, including clinicians like Chu, who specializes in primary care and infectious disease at Massachusetts General Hospital, the pandemic has forced difficult conversations about the limits of medical tests. It has also revealed the catastrophic harms of failing to recognize those limits.
“People think a positive test equals disease and a negative test equals not disease,” said Deborah Korenstein, M.D., who heads the general medicine division at Memorial Sloan Kettering Cancer Center in New York City. “We’ve seen the damage of that in so many ways with COVID.”
National COVID-19 test shortages have emphasized testing’s critical role in containing and mitigating the pandemic, but these inconvenient truths remain: A test result is rarely a definitive answer, but instead a single clue at one point in time, to be appraised alongside other clues like symptoms and exposure to those with confirmed cases. The result itself may be falsely positive or negative, or may show an abnormality that doesn’t matter. And even an accurate, meaningful test result is useless (or worse) unless it’s acted on appropriately.
These lessons are not unique to COVID-19.
Last year, David Albanese logged in to the online patient portal for his primary care doctor’s office and discovered that his routine screening test for the hepatitis C virus showed a positive result.
“I never considered myself somebody who’s in a high-risk category,” said the 34-year-old Boston-area college administrator and adjunct history professor. “But I just know that for a couple of days, I was really, really anxious about this test. I didn’t know if I should be behaving differently based on it.”
Within days, a confirmatory test showed Albanese did not actually have the potentially severe yet curable liver infection. Still, the memory of that false positive result gave him a new perspective on testing writ large. He had been skeptical of recommendations shifting breast cancer screening to older ages to reduce the psychological toll of false positives, but he said they made more sense after his own testing drama.
“‘Isn’t it better to do the screening regardless?’” he said he used to think. “Now, I realize it is a little more complicated.”
These false positives are especially common for screening tests like hepatitis C antibody tests and mammograms that look for medical problems in healthy people without symptoms. They are designed to cast a wide net that catches more people with the disease, known as the test’s sensitivity, but also risks catching some without it, which lowers what is known as the test’s specificity.
Though some degree of uncertainty is inherent in all medical decisions, clinicians often fail to share this with patients because it’s complicated to explain and unsettling and leaves doctors vulnerable to seeming uninformed, said Korenstein. What’s more, doctors are trained to seek definitive answers and can themselves struggle to think in probabilities.
“High-tech diagnostic testing has led to this mirage of certainty,” said Korenstein. “Back in the day before there were MRIs and what not, I think, doctors were more cognizant of how often they were uncertain.”
Enter COVID-19. Coupled with genuine uncertainty about an emerging disease and a political environment that has sown misinformation and rendered science partisan, the nuances of testing are too often lost at a time when they are particularly crucial to convey.
Jasmine Marcelin, M.D., who specializes in infectious disease at the University of Nebraska Medical Center, was concerned to see Nebraskans tested at statewide facilities get “inconsistent results without a lot of guidance or explanation about what these results might mean.” When she offers COVID testing, she said, she approaches it as she does any other medical decision, starting with a simple question: “What do you want to learn from this test?”
To answer this, it helps to know something about how coronavirus tests work and how well they do their jobs.
Many of the available tests are meant to tell you whether you’re infected right now. For example, polymerase chain reaction tests like the one Chu’s patient received detect small traces of genetic material from the virus. But by some estimates, those tests have a false negative rate of up to 30%, meaning three out of 10 people who truly have the infection will test negative. This rate also varies based on who collects the sample, from which part of the body and when in the course of a possible infection.
Antigen tests look for viral proteins and are faster to analyze than the PCR, but also less accurate.
To know if you’ve already had COVID-19, the closest you can get is the COVID-19 antibody test. But the too-common interpretation is black and white: I had COVID-19, or I didn’t. Here, again, the reality is more nuanced. The test checks your blood for antibodies—your immune system’s soldiers in the fight against the coronavirus. A negative antibody test could mean you were never infected with SARS-CoV-2, or it could mean that you’re currently infected but haven’t yet built up that army, or that these defenses have already faded away.
A positive test, on the other hand, may have mistakenly detected antibodies to another, similar-looking virus. And, even if the test correctly shows you had COVID-19, it’s not yet clear whether this means you’re protected from reinfection.
Yet, these shades of gray are difficult to internalize. Roy Avellaneda, the 49-year-old president of the Chelsea City Council, got the antibody test out of curiosity and could not help but see his positive result as what he called an immunity pass. “I can act a little bit cavalier with it now,” he said. “Yes, I’ll continue to wear a mask and so forth, but the fear is gone.”
Korenstein said that’s a common though worrisome reaction. “It’s really hard to expect the public to have a more nuanced understanding when even doctors don’t,” she said.
Some of the uncertainty around COVID-19 testing has abated as researchers learn more about the new disease. Early in the pandemic, healthcare providers retested patients with confirmed cases, looking for a negative PCR test to prove they were no longer infectious. But soon, epidemiologists discovered that a COVID-19 patient rarely infected others 10 or more days after first developing symptoms (or 20, in severe cases), even if the PCR test was picking up traces of the—presumably dead—virus weeks or even months after initial infection. So the Centers for Disease Control and Prevention and health systems adjusted their policies to clear patients on the basis of time rather than a negative test.
But while the desire for certainty in coronavirus testing is magnified by the rampant uncertainty in other facets of pandemic life, this is simply not something most medical tests can provide.