Not all coronavirus tests are the same: What to know about the different types
When it comes to diagnosing COVID-19, not all tests are created equal.
Detroit Lions Quarterback Matthew Stafford learned that when he was sidelined in early August by a false positive.
So did Ohio Gov. Mike DeWine, when a false-positive test kept him from meeting President Donald Trump during a visit to the state earlier this month.
While false positives can be inconvenient and lead to unnecessary quarantines and contact tracing, it's the false negatives that are of more concern to health officials, who say people who have the virus and get a negative coronavirus test could unknowingly spread disease.
"Some of these people that are asymptomatic; they may also be the super spreaders," said Dr. Daniel Ortiz, associate director of the microbiology and molecular pathology laboratory at Beaumont Health. "They might be the ones that are shedding high viral loads of a highly infectious virus and they end up infecting quite a lot of people."
As schools reopen and some require staff to test negative before returning to work, it's important to know what kind of test you're getting to determine whether your results are likely to be accurate, said Dr. Richard Zarbo, Henry Ford Health System's chairman of pathology and laboratory medicine.
And that's especially true if you're asymptomatic.
"There have been numerous tests, and the companies and laboratories who have developed the tests have received FDA (U.S. Food & Drug Administration) approval under their emergency use authorization," Zarbo said. "That doesn't mean that they're all equal, but they have been validated and accepted as tests."
For diagnostic testing, there are two primary types:
Molecular-based PCR (polymerase chain reaction) tests: These can detect the genetic material of novel coronavirus. Most samples for PCR tests are taken from the nose or the throat, although there are now a few saliva tests on the market. Some PCR tests can be processed quickly on-site. Others are processed in a lab with results returned in hours or several days.
Dr. Robert H. Shmerling, senior faculty editor for Harvard Health Publishing, wrote that the rate of false-negative molecular tests ranges from 2% to 37%. But false positives are more rare, and are likely in 5% of cases or less.
Antigen tests: These tend to be rapid tests that can detect proteins on the surface of the virus using a nasal swab or throat swab. Although a positive antigen test is usually considered accurate, a negative antigen test might need to be confirmed using a molecular PCR test.
Shmerling reports that antigen tests can have false-negative rates as high as 50%, but "because antigen testing is quicker, less expensive and requires less complex technology to perform than molecular testing, some experts recommend repeated antigen testing as a reasonable strategy."
But no test is 100% accurate, Ortiz said.
"There's no test out there that's 100% sensitive and 100% specific," he said.
"I think what a lot of people are going through now is that they just want to test for assurance. It's definitely important to realize what kind of tests you're getting and in what setting.
"A negative test today only means you're negative today. It doesn't mean a day from now or two days from now, you might actually reach that threshold of being positive."
Of the options on the market, one form rises above the rest in terms of accuracy, Zarbo said.
"If you're looking for a diagnostic test to diagnose whether you have COVID or not, the best test, the gold standard, is a PCR-based test out of a hospital-based laboratory or a reference laboratory," Zarbo said.
Several factors can affect the accuracy of any COVID-19 test. Among them:
When you test: Early on in a coronavirus infection, a person might not have enough particles of the virus to be collected on a swab and be detectable in a test. If you test too soon, you might get a false-negative result. Those who are deep into a coronavirus infection and are starting to recover might also not have enough virus particles to get a positive test result.
"It's kind of like a bell curve," Ortiz explained. "When you first catch the infection, you have low levels of virus. Then, that slowly starts to go up toward a peak, and you're shedding a whole bunch of virus, and then that slowly goes back down and trails off until you can't detect the virus anymore."
Sampling well: Another variable is how thoroughly the sample is collected. Did that swab get all the way to the back of your nose and swirl around enough to collect an ample amount of virus particles? If not, that too, could give a false-negative result in a person who actually has novel coronavirus.
Although it can be uncomfortable, "sampling using a nasopharyngeal swab has the most sensitivity, especially if you're asymptomatic," Zarbo said. The swab is pushed to the very back of the nasal cavity to collect a sample for testing.
"People like to use the nasal swabs" that only go as far as the front of their noses, he said. "It's easier to swab your nose than it is to have a very long, thin swab tickle the back of your brain essentially. People gag at that. But that's the most sensitive."
And though the FDA has authorized the use of some saliva-based PCR tests, which are less invasive, Zarbo said Henry Ford tried two studies to validate saliva testing in asymptomatic people.
They showed that it's difficult to collect a good sample, especially if patients are dehydrated because they're not creating enough saliva. The saliva tends to be too stringy and clogs machines used in the lab to process the test.
"Most of the samples were inadequate," Zarbo said, adding that the saliva tests also are far less sensitive than the nasopharyngeal swab PCR tests.
"We can detect down to 250 viral particles with a nasopharyngeal swab," he said. "You need at least 6,000 viral particles to 12,000 viral particles to be detected by that saliva test. So that's 25 to 50 times less sensitive.
"Why would you use that test if you're at a good chance of having a false negative?"
Contamination and other gaffes: Then there's the risk that the sample could have been contaminated either in the lab or during the collection process, which could yield a false-positive result.
That's what happened in the NFL when 77 players and staff got false-positive test results Aug. 22. The New Jersey-based BioReference Laboratories released a statement Monday saying its own investigation revealed that the “likely false positive results were caused by an isolated contamination during the testing preparation,” USA Today reported.
Test results also can be skewed if samples are kept at the wrong temperature before they are processed in a lab or if the reagent chemicals used to process the tests don't work properly, according to the FDA.
Zarbo said there's another factor that can play into false-positive results.
"When people bring up a false positive on any test, I always think there could have been a clerical error registering the patient and identifying the sample," he said. So I think misidentification is another issue."
While the idea of a rapid test result — whether it's an antigen or a PCR-based rapid test — can be enticing, Zarbo said accuracy can suffer.
The Abbott point-of-care tests that can give people results in about 15 minutes are "a lesser version of PCR," he explained. "The problem with that one is that the sensitivity of that is lacking compared to the PCR-based testing that we use in the large laboratories. Testing with Abbott has had anywhere from 15% to 50% false negatives."
The FDA issued an alert in May about the rapid Abbott tests, saying, they "may return false-negative results," and suggesting that "negative results may need to be confirmed with a high-sensitivity authorized molecular test."
"To us, that was an unacceptable false-negative rate," Zarbo said. "So, you can be fast, but you have to be accurate."
On Wednesday, the FDA gave emergency use authorization for a new Abbott antigen rapid test. Called the BinaxNOW™ COVID-19 Ag Card rapid test, it promises results in 15 minutes. The test is the size of a credit card, and costs $5.
It will connect to a phone app called Navica, where users can get their results and will also provide a temporary digital health pass that is renewed each time a person is tested with the date of the test result.
Abbott says its new test has a sensitivity of 97.1% and a specificity of 98.5% when administered by a health care provider on symptomatic people within seven days of symptom onset.
Ortiz said whether a person has symptoms when they're tested really matters.
"If you were to go to your local clinic, and to take like an antigen test, and you were negative, can you be assured that you were negative?" he said. "It all depends. So are you symptomatic or are you asymptomatic? That's a huge difference ... when you're testing with the antigen test. Typically what the antigen test is preferred is for people that are symptomatic. When it performs the best is usually within five days of your symptoms starting."
Since the pandemic began, both Henry Ford and Beaumont have each processed more than 100,000 COVID-19 tests in their labs, Zarbo and Ortiz said.
But, Ortiz said, testing is not a panacea.
The accuracy isn't where it needs to be — there's still a shortage of testing supplies like pipettes, swabs and reagents, he said. Those two factors alone suggest Michigan is nowhere near a point when it can test everybody — symptomatic or not — several times a month as a surveillance mitigation strategy that might allow the state's economy to fully reopen.
"When you talk about contact tracing and repeat testing and testing people multiple times per week or something like that, we haven't even tested the entire population once. So how are we supposed to ramp up and test that much where we can do it multiple times for the entire population? I just don't see it being feasible."
Contact Kristen Jordan Shamus: email@example.com. Follow her on Twitter @kristenshamus.