What’s going on?
The clamor for long-delayed coronavirus testing is teaching a basic lesson about how all medical tests work: No test is 100 percent accurate. Some test results may incorrectly say that a person has a condition, but they don’t. That’s a false positive. Other tests may incorrectly say someone doesn’t have a condition, but they do. That’s a false negative, and for covid-19, the illness caused by the coronavirus, at this stage of the outbreak, experts are more worried about this type of inaccuracy.
When a new test is rapidly created and deployed, its accuracy is often not fully known. The test is developed under controlled lab conditions, but it is used on samples taken, transported and performed by people in the real world — all of which increase the likelihood of errors.
The novel coronavirus tests use a swab to take a sample from the back of a person’s nose or throat. The swab is then transported to a lab that isolates and detects genetic material from the coronavirus.
Any medical test has two important qualities: sensitivity and specificity. The tests are proven to be “sensitive” in laboratory conditions — in this case, a technical measure of the smallest amount of virus they can detect. The tests must also be “specific” — for example, ensuring they do not mistake other pathogens, such as the common cold coronaviruses, for the new SARS-CoV-2.
The genetic tests being used are typically very sensitive and specific under lab conditions, but in the real world, how the swab was done and the stage of illness the person was in can make a big difference. To complicate the situation, there isn’t one test — many different tests are now being used by commercial laboratories, hospital labs and the Centers for Disease Control and Prevention. And the interpretation of the results will depend on not just the test, but other external factors, such as how widely the disease has already spread and laboratory practices.
“If it’s positive … you absolutely can make a [clinical] decision. If it’s negative, you may be early on in the infection and the viral load may be so low you don’t get it,” Anthony S. Fauci, the director of the National Institute for Allergy and Infectious Disease, said in a Q&A with JAMA.
A Cleveland Clinic researcher said the test developed by his hospital system is highly sensitive and specific in the laboratory, returning no false-negative results. But he acknowledged those numbers won’t exactly represent how the test will perform in the real world. Another researcher said anecdotal reports peg the genetic coronavirus tests being used in the United States at about 85 percent sensitive. That means that for someone who has the virus, there’s a 15 percent chance they test negative. A critical-care blog, EMCrit, estimated that the genetic tests are about 75 percent sensitive and suggests that a single negative swab doesn’t rule out the disease.
Documentation for the test approved for New York’s state lab explains the possibility of a false positive or false negative, emphasizing that a negative test doesn’t rule out infection.
“A negative result does not rule out COVID-19 and should not be used as the sole basis for treatment or patient management decisions,” according to the fact sheet for health-care providers. “When diagnostic testing is negative, the possibility of a false negative result should be considered in the context of a patient’s recent exposures and the presence of clinical signs and symptoms consistent with COVID-19."
People who test negative may find it confusing and hard to believe the result.
Laura Frazelle, a 34-year-old violinist from Virginia, found out her test was negative after a week in the hospital and 10 days after being tested. She was grateful to be home, but even more worried about covid-19 given the severity of the unspecified illness she had just survived.
″My final diagnosis was pneumonia in both lungs due to an unspecified infectious organism,” Frazelle said. “Yes, it’s a pretty big coincidence that I happened to be hospitalized for an infection that has all of the symptoms of a current global pandemic.”
There are a number of reasons a test might be negative when a person is sick with the coronavirus. It might be too early in the illness, when the amount of virus in the airway is still small. It could be a problem with how the swab was done. Different types of swab collection — the back of the nose, the throat, the outer nose — may also have different levels of accuracy, an issue doctors are actively debating given limited evidence. Then, there could be issues with the handling or transport of the swab. There could be laboratory error.
“Let’s say you had a swab that wasn’t obtained very well. … The person has the virus. But the nurse or physician just barely puts it into the nose because the person is backing up — we would not have a good specimen, so it could create a false negative in the test,” said Gary Procop, director of molecular microbiology, virology, mycology and parasitology at the Cleveland Clinic.
Early reports suggested that the genetic test used in China — which was a different design than the one developed by the U.S. CDC — was not very sensitive, meaning many cases were missed. There have been reports of sick people who had to be swabbed multiple times to test positive in China. Chinese clinicians used CT imaging scans of people’s lungs to diagnose the disease because those were found to be more sensitive — they missed fewer cases.
But the experience in the United States appears to be different, so far. Jeffrey P. Kanne, chief of thoracic imaging at the University of Wisconsin School of Medicine and Public Health, said that U.S. experts are not currently recommending CT scans to diagnose patients without the genetic test.
“A normal CT scan doesn’t exclude covid-19, and an abnormal may support it — but is not specific enough to avoid testing,” Kanne said.
To complicate matters, increasingly, Kanne said, he is hearing reports of cases that test positive for the virus but look on the medical scan more like a bacterial pneumonia, suggesting patients could have two diseases at once.
As flu cases are tapering off and covid-19 circulates in the general population, there is also a rising probability that anyone with symptoms has the coronavirus. That means that even those who test negative will increasingly still be counseled to act like they have it and stay in self-isolation.
Demetre Daskalakis, deputy commissioner for the division of disease control of the New York City Department of Health and Mental Hygiene, said he recently told a patient with mild symptoms who sought a test and received a negative result to act like he had it.
“The pretest probability if you have fever or cough in a pandemic — if you have a fever, cough and shortness of breath, it’s covid-19. Even if the test is negative,” Daskalakis said.
The test is also only true for a single point in time. A person who sought testing for symptoms caused by another pathogen could easily contract the coronavirus in the time between being swabbed and receiving the result. Or if they are early in the illness, they might test negative because the virus hadn’t multiplied enough.
“The test is a screening tool. One has to think of it in terms of probabilities, meaning a positive test is a pretty darn high probability of the virus. If you have a positive test, for all purposes, we consider a person infected and potentially contagious,” said Michael Z. Lin, associate professor of neurobiology and bioengineering at Stanford University. “A negative test is harder to interpret.”
If you’re a person with symptoms who hasn’t been in close contact with a confirmed case, Lin said, a negative is probably a true negative — although as the virus becomes more widespread in the community, it will become more difficult to know whether a person has been exposed. But if a person has been in contact with confirmed cases or is a health-care worker with repeated exposure, a negative test would indicate only that the swab taken at that point in time was negative, not that the person isn’t infected.
“Everybody thinks the lab test is always right. When we design tests, they [often] have a 95 percent sensitivity, a 95 percent specificity. That means 5 percent of the time, you’re wrong. That’s just a structure of testing,” Procop said. “You only want to test people you really do believe have the disease, and in this case, people you’re going to act on. If it’s an otherwise healthy, young person, you’re going to say go home and isolate yourself.”
Laurie McGinley contributed to this story.