clock menu more-arrow no yes mobile
Vehicles wait in line for Covid-19 testing at Dodger Stadium in Los Angeles on July 7, 2020.
Sarah Reingewirtz/MediaNews Group/Los Angeles Daily News via Getty Images

Filed under:

Should we be testing fewer people to stop the spread of Covid-19?

Smarter testing could help save the US pandemic response.

In late June, an Austin, Texas, man with a runny nose and sore throat got a Covid-19 test and was told it could take up to 10 days for his results to come back. While he waited, he shrugged off his symptoms as a cold and continued with his plans, which included attending a wedding outside Dallas.

Several days after the wedding, however, he felt much worse, with shortness of breath and a cough. So he went to the ER, where his physician, Natasha Kathuria, ordered a rapid Covid-19 test. It came back positive.

“If he had a test turnaround of 24 to 48 hours, he would have had a sustained sense of urgency, likely quarantined, and avoided infecting up to 150 people at a wedding,” Kathuria, who is also on the board of Global Outreach Doctors, a humanitarian nonprofit, tells Vox. Though Kathuria isn’t sure whether the patient infected anyone at the wedding, she says four or five of his friends have tested positive since. (It’s not clear if they caught it at the wedding or if they had been infected earlier, as “they had also been socializing quite a bit in the same group,“ she says.)

Although some Covid-19 results can be delivered within hours, 10-day wait times are now not unusual for results from the most common test — the kind that uses polymerase chain reaction (PCR) to look for an active infection — when patients who are not isolating can go on to infect others. While testing failures have been a blight on the US response since the beginning of the pandemic, the latest delays reveal a strikingly uneven system that hasn’t been able to scale up to meet spiking demand, stymying efforts to stop the virus’s spread.

It’s not just the new people with symptoms — and those with known exposure to the virus — who are stretching testing capacity, but also people who want assurance that they won’t infect others before traveling, socializing, or going back to work or school. And with the school year starting soon, the demand for testing is set to surge even more.

“There is a continuing, insatiable demand for testing that is expanding, from symptomatic patients to anyone interested in having a test performed,” Gary Procop, medical director of clinical virology at Cleveland Clinic and a board member of the American Society for Clinical Pathology, tells Vox.

Quest Diagnostics, which has run about one in five US Covid-19 tests, for example, currently has an average wait time of a week or more for most people, with some waiting up to two weeks, it said in a statement.

While the federal government and others have been focused on the number of tests performed and encouraging more people to get tested, the backlog has been piling up.

If people at a high risk for spreading the disease — such as essential workers — are facing the same delays as someone at a low risk, it’s not a terribly efficient system and makes the prospects for getting control of the virus fairly grim. “The idea of just telling people to go get tested I think is the real challenge,” says Michael Osterholm, director of the Center for Infectious Disease Research and Policy at the University of Minnesota.

Waits of longer than a day, other experts agree, severely hinder America’s ability to stop the spread of the coronavirus in the US. Delays “really undercut the value of testing, because you do the testing to find out who’s carrying the virus and then quickly get them isolated so they don’t spread it around,” Francis Collins, director of the National Institutes of Health, said on Meet the Press July 19.

Indeed, a paper, published July 16 in The Lancet Global Health, argues that if test results were provided the same day — and comprehensive contact tracing happened right away — about 80 percent of new transmissions could be prevented, effectively stopping the spread of the virus.

But if test results take a week — and even if contact tracing is quick and effective — we’re only stopping about 5 percent of onward transmissions, the researchers concluded. They didn’t even bother to extrapolate to longer wait times than a week, which plenty of people in the US are experiencing.

The ideal turnaround for test results would be no longer than 24 hours and preferably less, say experts. And many hospitals, clinics, and academic institutions can meet this time frame. But if results take longer than three days, as they do in many communities, particularly underprivileged, hot spot areas, “they’re totally useless,” says David Lubarsky, a physician and CEO of UC Davis Health.

Why are these delays getting so bad? “We don’t have enough test kits to test everybody, or enough labs or enough machines or enough trained personnel,” says Lubarsky.

Number of Covid-19 tests performed each day in the US.
Our World in Data

But it’s also about how tests, and results, are or aren’t getting prioritized.

Hospitalized patients typically get rapid results, often within a matter of hours. But when people in the community get a test — whether because they feel sick or because they are hoping to go on vacation — labs often aren’t being told whose tests to analyze first. So that leads everyone’s results to get delayed, including those who are most likely to be spreading the virus.

Though experts are adamant that community transmission must come down to alleviate some of the pressure on the testing system, they’re also calling on the government to approve rapid tests — and offer clearer guidance on who needs their results back first. Let’s dive in.

Labs are taking so long to process Covid-19 tests because the US still doesn’t have enough supplies

To understand what’s behind the disastrous delays in Covid-19 test results, it helps to look first at the labs, which are struggling to maintain adequate levels of basic testing supplies, much as they did earlier in the pandemic.

The American Clinical Laboratories Association, whose members include Quest, LabCorp, BioReference, the Mayo Clinic, and others, says labs are facing high demand for key materials for the testing process, from the test kits themselves to essential chemicals, like reagents, and even the PCR machines used to run the tests.

And the near future doesn’t look much better. “The global supply chain remains constrained,” Louise Serio, a spokesperson for the association, wrote to Vox in an email. As production and international trade remain slow, and demand continues to surge, many components remain hard to come by. “We anticipate continued shortages of the supplies and equipment.”

Other labs are finding basic plastic components, like pipette tips and plates, in short supply, Procop says, for the same reasons. And deliveries for many materials “are not always consistent.”

Academic labs are also struggling. UC Davis Health, for example, has invested $3 million in the past few months to build up its testing capabilities and now has the capacity to run more than 2,000 tests per day, providing results in less than 12 hours. But the labs aren’t able to get enough reagents to run more than 250 tests per day. “So we’re at about 12 percent of our current capacity,” Lubarsky says. “It’s just not right.”

As one Harvard public health expert put it recently in Time: “America’s testing infrastructure is collapsing.”

EMT Luis Chavez secures Covid-19 tests for delivery to a lab for processing on July 30, 2020, in Encino, California.
Brian van der Brug/Los Angeles Times via Getty Images

Delayed results are especially terrible for this virus

One of the biggest challenges in containing this coronavirus is that it frequently spreads before people develop any symptoms. In fact, an estimated 40 percent of people who get the virus may catch it from someone without symptoms. And others might be carrying — and spreading — the virus without ever getting sick. Which all makes it very challenging for people to make decisions, like avoiding all contact with others while waiting for test results, based on how they are feeling.

This presymptomatic and asymptomatic spread makes catching early cases particularly difficult, and essential, in high-risk settings like an assisted living facility. Lubarsky points out that delays in getting test results to prevent or stop outbreaks in these settings are especially harmful. If a facility can pinpoint infected individuals within a day of their test, they can quickly isolate them, find their contacts, and prevent much further spread. But, says Lubarsky, if results are trickling in over several days or a week, the virus has likely spread to other people, and the old results are, as he says, “absolutely useless.”

The long delays can also disincentivize even those with symptoms or with a known Covid-19 exposure from getting tested, says Osterholm. If people know they might have to wait more than a week for results that may no longer be relevant, they are more likely to figure, “why should I go get tested?” he says.

People might be especially reluctant to go in for testing if they know they will be in this limbo for a week or more. “One of the greatest impediments to viral containment is human impatience,” Lubarsky says. If you add to that people’s work realities and behavior, substantial testing lags “are just not acceptable and will not contain spread.”

The delays are also a reminder of the disparities between those who can more easily quarantine while waiting for results and those who cannot. For example, if you work a job that requires you to be physically present and doesn’t offer paid leave, it could be particularly hard to decide to miss work for 10 days as a precaution while you wait.

Additionally, many essential workers in relatively low-wage jobs who don’t get paid sick leave also regularly interface with the public. “So not only do they have a higher risk because of increased exposure, but they also increase the exposure to all of us,” Lubarsky says. “And while they might make a reasonable living, they don’t make the kind of living to take two weeks [off] waiting to find out if they’re positive.”

This is one of the big factors Lubarsky sees as contributing to the disproportionate spread of Covid-19 in underserved communities, including agricultural workers he sees in the Sacramento area. It’s also showing up in the preliminary scientific literature. One early report, cited in a recent Health Affairs article, found that in a mixed-income San Francisco neighborhood (the Mission District), Latinx people made up 95 percent of positive Covid-19 tests — and 90 percent of those who tested positive were unable to work from home.

As Bill Gates, the Microsoft co-founder and co-chair of the Bill and Melinda Gates Foundation, put it in a Tuesday CNBC interview, “You need to make sure that low-income communities that are most at risk, that they’re getting those results back within 24 hours.”

On the other end of the spectrum, many universities are proposing to test their students and staff on a regular basis with quick-turnaround on-site processes. Harvard University, for example, is already testing faculty and plans to test all residential students and staff every three days starting in the fall semester. MIT is planning to test students living on campus with a similar frequency — twice a week — providing results within 24 hours. Both can do so because their tests can be processed at the Broad Institute, a biomedical research center that converted its genomics facility into a test processing center.

“I think it’s important to recognize the disconnect between the turnaround time and testing in different places in our community,” Sarah Fortune, an immunologist at Harvard T.H. Chan School of Public Health, said on a call with reporters this week. Those working in labs on campus are already getting at least weekly tests, results for which they get back in less than a day. But if someone else in the Cambridge area were to get a standard test, they wouldn’t see results for upward of a week, she noted. “So there’s an enormous discrepancy.”

Testing delays are hindering contact tracing efforts

As people wait additional days and weeks before getting their Covid-19 test results back, their memories naturally get hazier about whom they could have spread the virus to. And because mobile phone-based coronavirus tracking is not widely used in the US, tracers are still relying on people to tell them whom they remember being in contact with — a task that gets harder with each passing day.

Even if we all had perfect memories, though, a delay in getting a positive result means that not only could that person be out infecting others, but also that those contacts could now also be spreading the virus.

For example, if the average symptom onset is about five days after infection, but people have the highest amount of virus in their system about a day before they start feeling sick, that means a delay of a week in getting results for one not-yet-symptomatic person could have sent the virus into two or so additional generations of patients. And with this virus’s exponential spread, if everyone infected goes on to infect an average of two other people, that means eight additional people now have the virus by the time contact tracing can even begin for the first individual.

And if the contacts face similar delays for their own test results, the new cases quickly pile up. (The number of new infections could be much higher if any of those people fails to physically distance and mask up in crowded places. This is also based on the assumption that people self-isolate once they start to feel sick.)

If that first person could have received their positive results in the same day, they could have reasonably been instructed to self-isolate, halting any forward transmission from them, and immediately informing any contacts to isolate and test, stopping spread there as well.

The authors of the new paper in The Lancet Global Health also map out how especially crucial rapid test results are for areas that are not using mobile app technology for tracing contacts.

The team found that viral spread could still be contained (reducing the average number of new infections from each person — known as the “R0” ratio — to below 1) with test result delays of about two and a half days, if 100 percent of the population were using a mobile contact tracing platform. If about half of a population were using the platform, test results could still come back after a day and a half. But with a conventional (person-based) contact tracing system, as the US is relying on, their model suggests test results would need to come back in less than a day to get the virus under control.

And not only is this not happening, but as cases spiral in certain areas, as they are in Florida and Texas, the budding contact tracing system gets overwhelmed, decreasing its ability to efficiently track every case. Or as Osterholm puts it: “It’s kind of like trying to plant pansies in a Category 5 hurricane — it’s not easy.”

How can we run more tests more quickly?

One thing that could improve the speed of getting results is, of course, using faster testing methods. The current PCR-based tests typically send tests to a lab for processing, which involves specialized supplies, machines, and personnel — not to mention the transport time and handling logistics. Some facilities, such as many hospitals, are able to do quick testing on site. (This is important, Lubarsky notes, because not only does that prevent the spread of Covid-19 from patients, but also every hour caring for someone who is potentially Covid-positive means another hour of staff using full PPE, yet another still-limited resource in some places.)

There are also other technologies we could be using to test for the coronavirus.

Many companies are at work on rapid at-home tests, which “would be tremendously helpful” if they are accurate enough, Osterholm says. (They would also need to be linked to the health department for tracking and communication back to the individual, he notes.)

There are concerns about robust accuracy in many rapid tests that are being developed — and some that have already been deployed. Some experts argue, though, that we shouldn’t, as they say, let “the perfect be the enemy of the good” in this case.

“We need the best means of detecting and containing the virus, not a perfect test no one can use,” Michael Mina, an epidemiologist at the Harvard T.H. Chan School of Public Health, and Laurence Kotlikoff, an economist at Boston University, asserted in an opinion piece in the New York Times. “Simple at-home tests for the coronavirus ... could be the key to expanding testing and impeding the spread of the pandemic.”

Others make the case that such rapid testing would also help us find more people who are infectious and let them know to isolate before they can spread the virus to others.

An interim step is “pooled” or “batch” PCR testing of samples. In mid-July, the Food and Drug Administration gave Quest emergency authorization to start using this process, in which some material from up to four tests is mixed together and run through the full PCR testing procedure. (LabCorp received a similar authorization in late July to pool up to five samples.) If the pooled result is negative, they were able to consolidate what would have been four or five analyses down to one. If the analysis picks up evidence of the virus, each one of the samples is then tested individually to determine which one (or ones) was positive.

With about 91.5 percent of tests coming back negative in the US right now, there is a good chance many batches will come back without signs of the virus, clearing all of the pooled individuals and freeing up that additional testing capacity for those who need it most.

A Quest Diagnostics aircraft is loaded at the Reading Regional Airport in Bern Township, Pennsylvania, to transport medical samples to testing laboratories on April 2, 2020.
MediaNews Group via Getty Images

To slow the spread of Covid-19, should we actually be testing fewer people right now?

Some experts say we need more, broader testing, but others insist officials should instead to be stricter — or, as they say, “smarter” — about who gets tested in the first place.

The Centers for Disease Control and Prevention currently provides rough guidance on the use of testing. But Osterholm and his colleagues put together a much more robust hierarchy for “smart testing” when resources are limited. They lay out the order of who should get tests when resources are lacking:

  1. Hospitalized patients with symptoms
  2. Symptomatic health care workers, first responders, essential workers, and those who work in high-risk facilities (like long-term care institutions or homeless shelters)
  3. Symptomatic people in the community
  4. People without symptoms who live in high-risk facilities

“That’s where we’re going to get the most bang for the buck,” Osterholm says of making sure testing resources are used for these groups and in this order. This strategy “would cut down on a lot of unnecessary tests,” he says, “so we could do more with the tests that we currently have, which would speed things up — less volume and more high-impact outcomes.” And, of note, when basic testing capabilities are limited, as they are now, they specifically recommend not testing in schools, most workplaces, or the general community.

California has instituted a statewide prioritization hierarchy, which has four distinct levels for testing. The first groups include hospitalized patients with Covid-19 symptoms and people who have been identified as a part of an outbreak. Only in the last group for testing — which are conducted if results in the state are taking less than 48 hours — do people who are asymptomatic but think they might have been infected and people receiving routine workplace testing get tests.

This is already playing out in the general community, which might be frustrating for many people but might be conserving testing resources for those with the very highest risk.

Some of the trouble nationwide, multiple experts say, is that many labs are not able to tell what category a person might fall into and thus are not always able to prioritize correctly.

Quest said that as of late July, it was stratifying “priority” and “other” patients, providing the former with a faster turnaround. But for the week of July 20, its average turnaround time for priority patients was still more than two days (versus one day the previous week). LabCorp also reported it was providing faster turnaround times for hospitalized patients.

UC Davis Health uses an algorithm to decide which tests to prioritize based on the risk of a person spreading the virus. “If there is a migrant worker who has suggestive symptoms, who lives in a multigenerational home, who’s about to get on a bus with 30 other workers, we want to know that now — as opposed to a 38-year-old executive who lives alone who has the sniffles,” Lubarsky says. “What’s gumming up the work is we’re testing all the worried well and don’t have a tiered system” for the country.

Those representing major testing labs want this sort of direction, too. “Now is the time to decide what kind of testing is needed, at what levels, and where to ensure we’re deploying these tools where they’re most needed,” says Serio of the ACLA. “For example, we recently received guidance [from the Department of Health and Human Services] to prioritize samples from nursing home patients in certain hot spots. Continued clear direction [like this] is critical to better manage demand.”

And without more widespread rules, the testing giant Quest has been requesting individual health care providers themselves be the gatekeepers for who gets tested and how many tests they send to the company’s laboratories “so that we can direct our capacity to patients most in need,” it said in a statement.

But, in general, the federal government seems to be moving in the opposite direction of more targeted testing. In late July, the FDA authorized the first test specifically to screen people without symptoms or any reason to suspect they might have been infected. The test from LabCorp, which has been in use for suspected Covid-19 cases since March, requires the same PCR process and equipment as other current tests.

The currently limited PCR resources, Osterholm and others say, should instead be deployed for the most actionable cases.

Otherwise, we will continue to overwhelm the testing system, and localities will continue to need to reinstitute shutdowns to keep the virus in check.

But the other big piece of the testing puzzle is actually using our other methods — masking, physical distancing, etc. — to decrease the spread of the virus so we don’t have to test as much. “We’ve got to drive these case numbers down,” Osterholm says. “If we only needed to test one-tenth the number of clinical cases, we can start matching supply with actual need. Right now, our caseload outstrips supply capacity.”

Procop suggests we could still be in the early days of this challenge, especially as some people call for regular testing of school students and staff, which could add a massive burden to testing laboratories.

All of this means, however, in absence of more organized guidance about who should be getting tests right now, it is up to people to make that decision on their own, Procop says. “Individuals visiting friends or going on vacation should not use precious testing resources and deny these to individuals in need,” he says. “They need to mask, respect social distancing as much as possible, and wash their hands frequently.”

This could help reduce wait times for those at real risk of spreading the virus. We have to stop the virus as much as we can now, Osterholm says, because this fall, “things are only going to get worse.”

Katherine Harmon Courage is a freelance science journalist and author of Cultured and Octopus! Find her on Twitter at @KHCourage.

World Politics

Can Democrats overcome their deep divisions over Gaza?


Why so many members of Congress are calling it quits


Hunter Biden’s new indictment, explained

View all stories in Politics