To get the US pandemic under control, a growing number of health and medical experts are making a clarion call for an additional testing approach to Covid-19.
What we need, they argue, are at-home rapid tests that look for antigens, proteins the live virus makes. These kits would allow anyone to test themselves for the coronavirus any time (and anywhere) for between $1 and $5, and get results in about 15 minutes. No doctors, labs, expensive machines, or special chemicals required.
“I see these [antigen] tests as a solution that’s literally sitting in front of us,” says Michael Mina, an epidemiologist at the Harvard T.H. Chan School of Public Health, who has led an effort to get these rapid tests authorized by the Food and Drug Administration. “We could be getting massive outbreaks down to nothing. ... It’s a no-brainer to me.”
Back in April, Deborah Birx, the White House coronavirus response coordinator, actually called for just this sort of solution in an interview on Meet the Press. “We have to have a breakthrough innovation in testing,” she said. “We have to be able to detect antigen.”
Several small companies have been developing these simple tests, and the conglomerate 3M is working with MIT on another one. (The new BinaxNOW test from Abbott, authorized August 26, is different — it requires a health professional to administer it, and the patient must have symptoms to get a prescription.)
And some of these at-home tests are ready to go. But they have been sitting on the shelf, unused, for months. Why?
The holdup is that rapid, over-the-counter antigen tests are not nearly as sensitive to the coronavirus as the molecular PCR (polymerase chain reaction) tests currently used for most Covid-19 testing. In fact, by PCR standards, some at-home antigen tests might catch only half or a third of people who test positive on a PCR test.
But that’s because they’re looking for different things in different ways. A PCR test, designed to diagnose an individual with infection, can catch very low levels of viral material before or after a person spreads the virus. An over-the-counter antigen test, meanwhile, relies on high levels of the virus to be able to detect it. These high levels, though, also happen to coincide with when a person is most likely to transmit the virus to others.
“The antigen test is maximally sensitive at the same time that you are maximally infectious,” says David Paltiel, a professor of public health and health policy at the Yale School of Public Health. “The antigen test is picking up the infections I want. Because I don’t care about infections, I care about infectiousness.”
Since at-home antigen tests would be fast and cheap, they could allow people to test themselves regularly. Currently, PCR testing is relatively expensive (about $100) and often experiences processing delays that hamstring efforts to stop transmission.
PCR tests’ sensitivity, however, is the current benchmark by which the FDA judges new tests for emergency authorization.
This pre-pandemic framework, Mina and others argue, is a problem as we look for the most effective tools to fight the coronavirus right now. So they are calling for a new way to evaluate these tests, one based in public health terms rather than individual diagnostic ones — in other words, widely accessible tests that are well tuned to find those who are most likely to spread the virus.
Let’s take a look at how and why this potential solution has gotten sidelined in the pandemic battle — and what might be done to get these tests off companies’ shelves and out to the public.
A test to find infectiousness, not infection
To understand the stalemate around antigen tests, it’s helpful to first understand what they are, how they differ from classic PCR tests, and the trajectory of coronavirus infections.
First: An antigen test looks for a particular protein from a live virus. (Not to be confused with an antibody test, which finds immune cells your body has made after mounting a defense against the virus.) These tests need a lot of viral material to generate a positive result.
Second: A PCR test looks for the virus’s genetic material — its RNA — making copies of itself until it reaches a detectable level. As a result, it has a fairly low (although not perfect) false-negative rate, or the proportion of time it would tell someone who has the virus that they don’t.
Finally: One of this new coronavirus’s superpowers is its ability to spread from people before they start to experience symptoms. In fact, people tend to carry the most live virus the day or so before they begin feeling sick, and the amount tends to quickly trail off in the several days after symptom onset.
So proponents of at-home antigen testing say that PCR tests, while useful in determining whether an individual is infected with Covid-19, are actually a poor tool in finding people who are most likely to spread the virus. That’s because PCR tests are so sensitive, they are excellent at picking up traces of the virus even after someone has beat it back and is no longer infectious.
Thomas Tsai, a health policy expert at the Harvard T.H. Chan School of Public health, compares the two types of testing this way: PCR tests are reactive tests, usually used for people who either have symptoms, a close contact with the virus, or other reason to think they are likely to have caught it. Rapid, widespread antigen tests are proactive tests, designed to find cases before they spread the virus more widely.
And as the Center for Evidence-Based Medicine noted earlier this month, “while an infectious stage may last a week or so because inactivated RNA degrades slowly over time, it may still be detected many weeks after infectiousness has dissipated.”
To look a little bit deeper at how this works, an August 7 preprint report which hasn’t been peer reviewed proposes that people with 10,000 copies of the virus detected in their airways are unlikely to spread the virus to someone else, even with sustained contact (according to their models, it would happen about 0.002 percent of the time). Once that viral load climbs to 10 million copies (which happens quickly, potentially in about a day), they have about a 40 percent chance of transmitting it to a close contact — and about 80 percent chance at 100 million viral copies (it’s worth noting that when they first reach this level they still might not have any symptoms).
By the time someone has had Covid-19 symptoms for several days (which might currently be about the time they receive results from a PCR test), however, the amount of virus in their system has likely waned to the point of actually being less transmissible. As the authors of the study note, “transmission after the first week of infection is quite rare.”
At that point, says Mina, it’s more likely that “they just have residual [virus] RNA hanging about.” So although many of these people would get a positive PCR test — and a negative at-home antigen test — they are much less likely to transmit the virus than someone who feels fine now but will develop symptoms in a day or two.
“The PCR is being fooled all the time,” Paltiel says. “The PCR is just picking up strands of viral junk.” If your goal is to slow an outbreak, he notes, this sort of data is not actually that useful. “Outbreak control is all about figuring out who’s actually spreading this stuff.”
This could also help explain why so many people continue to get positive PCR test results weeks after getting over most of their symptoms. One small JAMA study, for example, found that one in six former Covid-19 patients with no symptoms tested positive for the virus via PCR test four to 24 days after being discharged from the hospital. As an author of a commentary about that paper noted about these long-tail positive test results, “the clinical significance and infectivity are minimal. These PCR tests likely are responding to noninfective RNA fragments and do not represent detection of viable virus.”
Antigen tests, however, need a much higher level of the virus to register a positive result because they don’t rely on multiplying their target.
This, experts argue, make them well matched for finding people with Covid-19 at precisely their most infectious points rather than potentially long after symptoms have arrived and infectiousness has faded, as is more likely now with the long delays in PCR results and contact tracing. As Paltiel describes it, the concept is pretty simple: “The more viruses around your airway, the more likely you are to be a risk to others.”
According to an analysis from the June preprint (on which Mina was a co-author), this crucial time period is precisely where the antigen test is most sensitive: detecting viral load at around 10,000 copies, which is just before someone’s infectiousness skyrockets. PCR tests, on the other hand, can pick up about 1,000 copies of the virus, which gives them about a day’s head start at finding the virus. Although with a lag of more than a day in returning results, that benefit is lost.
To put it in other terms, these tests could be “superspreading detectors,” Mina says, locating many of those who would otherwise go on to unknowingly infect many other people before they know they have the virus.
“Their whole goal is to capture the 90 percent of people who might be transmitting the virus,” Mina says. So if they were to be evaluated in this population — those with high levels of the virus — he says, the sensitivity for these tests would soar from 30 or 50 percent (as measured against PCR) up to 95 percent. (They also have very low false-positive rates, in which they would tell someone they had the virus when they actually didn’t.)
Another selling point to antigen tests, proponents argue, is that they would give people their results much faster than PCR tests. Since delays of more than a day or so in getting PCR results back make containment and tracing much less effective, being able to have results back in 15 minutes or so would be a vast improvement.
Because these tests would be fast and affordable, people could also test frequently, which could help overcome worry about the tests being less sensitive than PCR — especially in the very small window early in an infection when a PCR test might catch the virus but an antigen test wouldn’t.
And such frequent testing would be far superior to the still relatively rare PCR testing that we’re currently doing, Paltiel says. “Of all the variables that we control, the frequency of screening is the most important,” he says. Even if the antigen-based test only caught half of all cases (regardless of infectiousness level), he says, if it’s “between a test that is missing 50 percent [but rapid and easy] and a test that is getting everyone that is so expensive or so cumbersome or has such a long turnaround time that it takes a week, I’ll take what’s behind door No. 1.”
In the June preprint study, the researchers found the sensitivity of a test made little difference in the rate at which it could bring viral transmission down. For example, running either of these tests, PCR or at-home antigen testing, on a population weekly could lower the spread of the virus by about 60 percent, they posit.
We can bring this pandemic under control within weeks— Ashish K. Jha (@ashishkjha) August 17, 2020
Without shutting down the economy
With technology we have now
But we won't
Because of @US_FDA backward thinking about tests
Watch @j_g_allen video. Then, don't get mad. Let's act.#dailyquicktest https://t.co/dWYCE8JpNA
So, Mina suggests, we shouldn’t just be asking what level of sensitivity a test has, but rather, what it’s sensitive to. PCR tests are great at finding traces of the coronavirus but, he says, “it’s going to have a 0 percent sensitivity to detect elephants.” And if what we want is to start finding elephants — or people who are most infectious — we should look for new testing options, he argues.
The FDA’s job is to help keep us safe, but it might lack a pandemic public health protocol
One of the FDA’s essential responsibilities is to make sure dangerous medications, ineffective vaccines, or misleading diagnostic tests don’t make it to US consumers. So it applies rigorous standards to make sure new products are reliable before they are allowed on the market.
And the FDA has applied this same rubric when it has evaluated new tests for the coronavirus this year.
“Tests are still being considered first and foremost as diagnostic tools,” Mina says. This is evident in how the FDA has required antigen tests to perform at PCR-levels in detecting traces of the virus in people’s systems, he notes. For example, a new test might need to detect the virus in 80 or 90 percent of emergency room patients who’ve received a positive PCR test.
So when you compare the numbers and see a test that catches half or a third of the positive cases a PCR test does, it makes sense for the FDA to worry about a huge number of Covid-19-positive people thinking they are negative.
And Mina says he gets that. “If I’m a doctor, which I am, and I have a patient in front of me, I would want the absolute best and most sensitive molecular test to make sure I don’t miss something,” he says.
But he is also an epidemiologist. “For public health, it’s totally different. It requires a really different type of thinking.” One that targets minimizing spread of the virus among people.
“Unfortunately, the FDA just doesn’t have that,” he says. “They don’t even have a lens with which to think about it.”
And so, he says, they’re still evaluating at-home antigen tests as individual medical diagnostic tools rather than “as a virus control tool at the population level.”
The FDA has already authorized four antigen-based tests, including ones from Quidel (in May), BD Veritor (in July), LuminaDX (in August), and the new one from Abbott. To meet the FDA’s current sensitivity standards, however, the first three of these use proprietary machines to read results. (Some of these machines are now being allocated by the Centers for Medicare & Medicaid Services to nursing homes around the country.)
But an over-the-counter antigen test, the sort that Mina and others are advocating for, simply can’t be authorized under the current standards.
This gets at the essential tension between the FDA’s individual-based medical diagnostic framework and the broader public health needs of the pandemic. And, says Tsai, “We really need to shift gears.”
Concerns remain about at-home antigen testing
Beyond the sensitivity question, the FDA and others also have hesitations about these sorts of tests being done outside of the health care setting.
Current Covid-19 tests require a health care professional to order them and provide the results, allowing for them to give patients medical advice. If people are taking these tests on their own at home or before entering a business, a nurse won’t be on hand, for example, to advise them to isolate if necessary or take other steps to prevent spread.
“Due to the lack of healthcare professional supervision, FDA believes it is important for over-the-counter non-lab diagnostic tests to have a low rate of false negative results,” Emma Spaulding, a spokesperson for the FDA wrote to Vox in an email. “For example, an individual with a false negative result from an OTC diagnostic test may be less likely to quarantine despite symptoms, putting others in the community at risk,”
This is one reason they are recommending such tests catch 90 percent of the PCR-positive cases, she says. They would lower this to 80 percent if the tests required a prescription and were done under the supervision of a health care worker (such as via telemedicine).
Others worry about individual compliance. Rebecca Lee Smith, an epidemiologist at the University of Illinois, notes that this sort of testing would probably be most readily adopted by people who are already putting in the most effort to avoid catching and spreading Covid-19, such as those regularly wearing masks and practicing physical distancing.
“But the people who aren’t concerned about the virus may not choose to do at-home testing — they’re also the people who are not taking the precautions. So the same people we would absolutely want to be checking for the virus, then, would be the people less likely to do it at home.”
She also has concerns about at-home tests being used to clear people for work, school, or other activities. If someone needed to produce a negative test result, they could theoretically get someone else to take the test for them but still gain entry to, say, their college classes for the day. That was a point of discussion at her university, where they eventually opted for supervised (rather than collected-at-home) saliva-based PCR testing. “We decided that it was more important that we know, for sure, that the person who checked in for testing is the person giving the sample,” she says.
Administering these tests out in the community as a way to enter a workplace or a restaurant, for example, could also help alleviate this issue — as well as the concern about those with higher chances of contracting the virus not taking the test.
The newly approved Abbott test, BinaxNOW, is a step toward more accessible antigen testing. It doesn’t require a machine to read the results — and still, according to data presented to the FDA, has a 97 percent specificity level compared with PCR tests. It is also $5 for one test, and the company says they will be able to ship 50 million of them a month this fall.
The downside is it is only available with a prescription, and it needs to be given by a health care professional or other trained individual (such as a pharmacist or workplace health specialist). The test requires a nasal swab and a small amount of specialized chemicals known as reagents. And it is only to be used for people who developed Covid-19 symptoms within the past week, making it inapplicable for broader population-wide screening.
With an over-the-counter test, we might also lose a lot of important public health disease surveillance data. If people are testing positive at home — or even in a public setting — and not following up with a health care worker or public health department, their case might not get counted or their contacts traced. “You would need to have a link back to public health,” Smith says. They would also want to get information on the test that was performed so they could properly interpret the results.
This is where our existing PCR testing capacity could run backup. Smith explains that a positive antigen test result could be a trigger to get a PCR test to confirm the infection. It could also possibly reduce the number of PCR tests that are run, potentially speeding up delivery of those results as well.
“In a perfect world,” Smith says, “we would have this cheap, at-home test that anybody could take, as frequently as they want, that would give them a quick result. If there is a reason — if they have symptoms, if they’ve been exposed, or the at-home test comes back positive — then they report for PCR.”
Tsai agrees PCR tests would still be important. “It’s really thinking about how we use the pros and cons of all the different tests in a more comprehensive strategy,” he says. “Let’s put these pieces together.”
How could at-home antigen tests get authorized?
With over-the-counter antigen tests at a regulatory stalemate in the FDA’s current authorization system, experts are spitballing alternative routes to getting these tests to the public.
First and foremost, says Mina, government involvement in authorizing new tests is essential to make sure that those that do reach the market work in the way they are supposed to.
One route for this, he proposes, is a new standard by which the FDA could authorize tests, setting a different benchmark for test sensitivity at levels of the virus that are most likely to be transmissible rather than at very low detectable levels. In other words, he notes, the FDA could adjust their language from looking at sensitivity to the virus broadly to “sensitivity during peak infection.”
The FDA could also reframe the way it characterizes these tests, he says. Instead of being evaluated as an individual diagnostic test, he says, “it would essentially be indicated as a transmission-detecting test ... [or] a public health diagnostic test, where the real reason of doing that diagnostic test is one of public health.”
Another option would be to designate these tests as surveillance tools rather than diagnostic tests. Under that category, they fall outside the FDA’s purview and would more likely be overseen by the CDC and local health departments.
This plan has a big hitch, though. Surveillance test results are aggregated and are not shared back with the individual taking the test. Diagnostic tests, on the other hand, are those which have results given back to the individual so that they might take a specific action. The latter is, of course, the goal here (to let people know, for example, if they need to start self-isolating).
“It’s this awful catch-22 that could easily be changed if there was a will,” Mina says. “There just doesn’t seem to be much of a will.” Or a regulatory framework for thinking about things differently — even in the midst of a pandemic, he says.
The FDA has recently conceded that they would consider authorizing a less sensitive test if it were part of a high-frequency testing plan, with each person being tested multiple times (which they call “serial testing”), Spaulding says.
This, to Mina and others, seems like a step in the right direction, and he is hoping the FDA will provide more detailed guidance on what this would look like in practice.
But the FDA also notes that it would require an application for this sort of serial testing to “include the capacity to manufacture a sufficient supply of tests with which to conduct multiple tests per person,” Spaulding says.
Mina suggests that this is an arbitrary ask. For other authorized tests, like those using PCR or machine-based antigen detection, the agency has not required makers to show manufacturing capacity — or even the supply chain to turn results around in a certain amount of time (which is a key aspect to slowing transmission of Covid-19). “They’re perfectly fine approving tests that might take a year to return results, but that’s a 100 percent useless test,” Mina says.
Mina also worries that as some of these companies with at-home antigen tests wait for the government to greenlight them, they will decide to make them more complex to meet the current (PCR-based) requirements for diagnostic viral detection.
He calls this going from the sort of “instant coffee model” of testing (where anyone can take a test, cheaply, pretty much anywhere) to the “Nespresso model” (where you need access to a specialized machine to get a result).
And slowdowns in rolling out these machine-based antigen tests are already happening. Both BD and Quidel, two companies making rapid antigen tests, are now facing supply chain issues as they try to fulfill orders for their machines and tests, the Wall Street Journal reported.
Tsai notes that public health officials in Florida have been advised to use the Quidel antigen tests only for symptomatic, older adults. Which, he says, makes sense in targeting those most at-risk for complications from the virus, but “in some ways also defeats the purpose of the frequent testing strategy.”
Mina has been challenging those who are still on the fence about rapid, at-home antigen testing to say whether, in hindsight, it would have been good to have five months ago.
“Almost all of them say, ‘yes, that would have been great.’ Compared to what has happened, of course it would have been great.” Now, he says, “we’re in an almost worse spot today, we’re hitting up against September and October, when coronaviruses can start spreading like crazy. We have to cut our losses and say, ‘okay, we didn’t do it five months ago, but we can do it today.’”
Like the old tree-planting aphorism: The best time to plant a tree may have been 20 years ago, but the second-best time is now. “We could potentially save ourselves from ourselves in the fall. But I don’t know if it’s going to happen, frankly,” he says.
Part of that reason, he notes, has to do with our general national mindset. “I think Americans, including policymakers, are having a hard time coming to terms with the idea that there’s actually something bad happening to us.”
Just adding this new type of testing on its own, however, won’t be enough to get us out of the pandemic. “We can’t test our way out of it,” Smith says. “We also need masks, we need distancing.”
But, she says, many presymptomatic and asymptomatic cases are not being detected with our current testing strategy. “And those people are able to spread the infection further without knowing it. If we can’t identify these cases, we are never going to get out of this.”
Katherine Harmon Courage is a freelance science journalist and author of Cultured and Octopus! Find her on Twitter at @KHCourage.
New goal: 25,000
In the spring, we launched a program asking readers for financial contributions to help keep Vox free for everyone, and last week, we set a goal of reaching 20,000 contributors. Well, you helped us blow past that. Today, we are extending that goal to 25,000. Millions turn to Vox each month to understand an increasingly chaotic world — from what is happening with the USPS to the coronavirus crisis to what is, quite possibly, the most consequential presidential election of our lifetimes. Even when the economy and the news advertising market recovers, your support will be a critical part of sustaining our resource-intensive work — and helping everyone make sense of an increasingly chaotic world. Contribute today from as little as $3.