Dr. Sara Cody, health officer of Santa Clara County, California, was tired of seeing the same thing over and over again. Her contact tracers were telling people exposed to COVID-19 that they needed to get tested, but when some went to testing sites, health care providers turned them away because they didn’t have any symptoms.

This posed a problem for Cody’s work. Knowing if a contact was infected would help her department keep an accurate count of her county’s coronavirus infection rate; also, if a contact tested positive, it’d spur a new round of contact tracing from her staff, to help stop any further transmission from that asymptomatic carrier.

Cody decided to issue a countywide health officer order in June requiring certain health care facilities to provide testing for all close contacts, and also all front-line workers, such as mass transit drivers and retail workers, whether or not they had symptoms.

Then last week, the U.S. Centers for Disease Control and Prevention quietly changed guidelines on its website to say that people without symptoms did not necessarily need to be tested, even if they had been in contact with someone who had COVID-19. Cody was confused. ”Was it because there isn’t enough testing capacity?” she initially wondered. But there was no such explanation from the agency.

The CDC was met with a degree of pushback that was notable in its intensity; several states flat-out said they would not follow the guidelines, including California, where Gov. Gavin Newsom said, “I don’t agree with the new CDC guidance. Period. Full stop.”

The controversy surrounding the CDC guideline change is all just a symptom of a deeper issue that has plagued America’s coronavirus response: Even though we have spent more than half a year battling a virus whose insidious hallmark is its ability to spread through those with no symptoms, the country has not yet articulated a coherent strategy to test these silent carriers.

“The fact that we’re this far into the pandemic and we’re still talking about how to do asymptomatic testing and going back and forth on this is a major part of the reason why we’re struggling to open schools and colleges, and why people are still dying in prisons,” said Dr. Ashish Jha, dean of the Brown University School of Public Health.

The lack of consistent asymptomatic testing guidelines means that from state to state, county to county, a hodgepodge of strategies are being used with varying standards, testing methods and levels of access. Decisions are being made sometimes by people who have been thrust into the role of public-health officer with no training — school principals and college deans, leaders of companies and daycares and churches, who are just trying to do right by the people they are responsible for.

It’s unfair to ask them to have to come up with their own testing strategies, or to have to navigate the maze of their local health authority’s often shifting recommendations. There may be pros and cons to various strategies experts have proposed, with variables to consider like testing technologies, supply chains and federal funding, but perhaps the more urgent need at this point is picking a plan and actually seeing it through.

To understand why, we need to start with a clearer understanding of the pivotal role asymptomatic testing plays in containing this virus, particularly in the absence of a vaccine.

Why Asymptomatic Testing Is Important

Let’s start with the most basic question: Why do we bother testing in the first place? There are, broadly speaking, two reasons to use a test. The first is as a clinical diagnostic; the other is as a public health tool. Both are important, but for different reasons.

Doctors use a clinical diagnostic like a strep test to tell whether a patient is sick with a disease that can be treated with particular medicines. “The purpose of the test is based around doing one thing when it’s negative and doing another thing when it’s positive,” said Dr. Patrick O’Carroll, head of health systems strengthening at the Task Force for Global Health, who previously worked at the CDC for 18 years.

From this perspective, it seems pointless for an asymptomatic person who might have COVID-19 to take a test, because there’s not going to be any difference in how they will be treated — there are no symptoms to medicate. You might have even heard your doctor say, “Don’t bother taking a test if you only have mild symptoms, because I’m not going to tell you do anything different besides drinking fluids, taking Tylenol and resting.”

But from a public health standpoint, testing asymptomatic people can yield actionable information. COVID-19 is unlike many other diseases, in which a patient’s peak contagiousness coincides with the height of their symptoms. With COVID-19, about 40% of patients do not show any symptoms or have such mild ones that it would never have occurred to them that they had been infected. In a recent study of 192 young people with suspected COVID-19 in Boston, only half who tested positive had a fever.

Furthermore, studies have shown that among patients who do develop symptoms, viral load, which correlates with a patient’s contagiousness, is highest right before or at the time when symptoms start appearing. Put together, these features have explained why the coronavirus has been able to spread so perniciously across the globe. It’s one sneaky virus.

If an asymptomatic person tests positive, public health officials can ask them to isolate from others and begin the process of contact tracing in order to break chains of transmission. In the bigger picture, it also helps them keep tabs on where the virus is spreading in their city. (This is what’s known as “surveillance” in public health parlance: They’re not spying on you. They’re tracking the virus.)

“Since the beginning, testing has been the foundation of our response, because it tells us who is positive, where they are, in which demographic, and what the patterns are,” said Dr. Umair Shah, executive director for Harris County Public Health in Texas.

Understanding population prevalence also helps guide public health actions. For example, said O’Carroll, “if testing shows that only 2% of the population is positive, I’m going to call all of those people, interview them, put all of the contacts in quarantine and really try to stamp it out. But if I find that 30% are positive, then I really don’t have the resources to interview and chase down thousands and thousands of people — that’s when spread is too high for contact tracing to be useful.”

When testing is restricted to symptomatic patients, health officials will only have limited signals about the extent of the virus’s spread, leaving them to operate partially blind.

Standards Vary From State to State

There are two categories of asymptomatic people to consider: The first includes those who had close contact with someone who has already tested positive for the virus. The second includes people who don’t have any reason to believe they have been exposed. The first group is a higher testing priority, because there’s a far greater chance that they have been infected and can be spreading the virus.

In an ideal world, if testing were abundant and cheap and results were fast, we would test everyone daily and catch all of the asymptomatic carriers. But when there aren’t enough tests to go around, public health officials need to triage.

In the earliest stages of the pandemic, when there were hardly any tests available across the country, public health officials had to limit tests to the most urgent need — people with severe symptoms in hospitals. As tests became more available, they started to widen the criteria, first to people with symptoms, then to asymptomatic people with known exposure. Finally, in some areas of the country, anyone who wanted a test could get one, whether or not they had symptoms.

But to this day, the decisions have been made piecemeal. I reached out to health departments around the country, and found that testing criteria still vary depending on where you live.

In Delaware, close contacts are asked to get tested once, at the end of their 14-day quarantine period. The state lets anyone get tested, whether or not they were exposed or have symptoms. Maryland recommends that people who suspect they’ve been exposed to the virus get a test, whether they are symptomatic or not. Arkansas says it works to facilitate testing for all close contacts of positive cases, and also tries to provide testing for anyone in the state who wants a test, asymptomatic or not.

But Oregon and Wisconsin don’t recommend testing for asymptomatic people who have not had close contact with a confirmed case. (Oregon makes an exception for people in a high-risk category, such as agricultural workers.)

Some states have more nuance to their recommendations. New Jersey said testing is available to all, but noted that if you are asymptomatic, testing is recommended if you are a front-line worker, if you were in a large crowd with difficulty social distancing, if you are a member of a vulnerable population or if you recently traveled somewhere with a high COVID-19 infection rate.

Effective Contact Tracing Can’t Happen Without Efficient Testing

Within each state, however, guidelines aren’t always followed consistently by test providers. Cody, the health officer in California, isn’t the only one whose contact tracers are unable to get asymptomatic people tested.

Rebecca Fischer, an assistant professor at Texas A&M University School of Public Health, said she’s seen the same thing happen in Brazos County. “We call them and say, ‘How did the test go? And they’ll say, ‘They sent me away because I don’t have symptoms.’ and we’ll say, ‘You need to go back and say the health department sent you,’ and often they get turned away again.” Sometimes, Fischer said, the health department would have to give the person a letter to verify that they needed a test.

“We get on the local news station and plead with test providers to help us facilitate widespread testing,” Fischer said.

It’s unclear why providers are turning down asymptomatic patients. It may be, in part, due to the perceived purpose of the test. Dr. Michael Hochman, a primary care doctor and director of the Gehr Family Center for Health Systems Science and Innovation at the University of Southern California, said he thinks the value of testing contacts without symptoms is “modest” and would rather make them stay home for 14 days instead of come into a clinic for a test, “which is bringing them together with other people, the opposite of what you want.”

He worries that a false negative could give patients a misguided sense of security and prompt contacts to leave quarantine before they’re supposed to. Hochman says he sometimes has patients calling who say they have potential exposure and want a test, but when he explains to them that regardless of the result, they still will need to quarantine, the patients often then decide they won’t bother with a test.

Cody countered that many people don’t always adhere to the 14-day guidelines. “We’re not doing legal orders, so there’s not going to be perfect compliance,” she said. Given the opportunity to test and find out that an asymptomatic contact is positive is always preferable, she said, because people are more likely to take precautions and isolate properly, particularly around family members.

Without a Clear National Strategy, Confusion Abounds

Into this already chaotic environment came the CDC’s guidance change on Aug. 24.

Normally, when the agency updates its guidances, it gives a heads-up to state and local health departments, so they can decide how to adjust their own recommendations or how to communicate to the public, said Chrissie Juliano, executive director of the Big Cities Health Coalition, which represents large metropolitan health departments.

“Usually at minimum, there’s a big tent call … and normally at the top of the call, they say, we’re going to update this.”

But this time, it didn’t happen. “It was buried in an email,” she said. “If you hadn’t clicked on it, you wouldn’t have known.”

Previously, the CDC recommended testing for all close contacts of people with known COVID-19 infection, specifically noting that “because of the potential for asymptomatic and pre-symptomatic transmission, it is important that contacts … be quickly identified and tested.” The new guidance, however, said, “If you have been in close contact … you do not necessarily need a test unless you are a vulnerable individual or your health care provider or state or local public health officials recommend you take one.”

The new guidance for asymptomatic people who had no known exposure conveyed a number of different messages, depending on which part of the website you read. On one hand, it said, “If you do not have COVID-19 symptoms and have not been in close contact with an infected person: You do not need a test.”

But farther down the page, the site also said, “If there is significant spread of the virus in your community, state or local public health officials may request to test more asymptomatic ‘healthy people.’”

In the absence of explanation or context, confusion ensued.

Calling the new guidelines “vexing and hard to interpret,” Dr. Jeff Duchin, health officer for public health for Seattle & King County said in a statement that “testing asymptomatic close contacts of COVID-19 cases is important to identify cases and interrupt transmission and we intend to continue to do that pending additional information that would lead us to reconsider.”

When I asked the CDC to explain the change in guidance, it didn’t respond, instead pointing me toward the Department of Health and Human Services.

HHS sent me a statement from Adm. Brett Giroir, the federal testing czar, saying that the updated guidance “places an emphasis on testing individuals with symptomatic illness, those with a significant exposure or for vulnerable populations, including residents and staff in nursing homes or long term care facilities, critical infrastructure workers, healthcare workers and first responders, and those individuals (who may be asymptomatic) when prioritized by public health officials.”

The revised guidance did not appear to be generated internally by the CDC. Giroir later told reporters that the recommendations were approved by members of the White House coronavirus task force, saying, “We all worked together to make sure that there was absolute consensus that reflected the best possible evidence.” Dr. Anthony Fauci, however, said he was undergoing surgery and was not part of the discussion.

A few days later, CDC director Dr. Robert Redfield verbally softened the changes, saying that testing “may be considered” for asymptomatic contacts, though the guidelines online were not changed.

“Ultimately, it may not actually be a huge change,” said Juliano, but in practice it means that the federal government “is really pushing the decision down to states and local.”

“It means when public health says you should get tested, someone could say, ‘well, the CDC says it’s not necessary.’ It leads to public confusion, and you’re really putting state and local in a line of fire that’s not necessary.”

Use the Right Test for the Right Situation

Now that we’ve talked about the reasons it’s important to do asymptomatic testing, it’s time to think about resources. In recent months, many experts have been advocating that different types of tests be used for different purposes, in order to optimize available supplies and avoid testing delays.

The idea goes like this: We should save the most sensitive tests — known as PCR tests — for diagnostic purposes, when we need to be absolutely sure that a patient has COVID-19, because we’re going to be treating them or asking them to isolate, based on the results. So these tests should be used for people with COVID-19 symptoms and people who were known to be exposed to the virus.

But for public health purposes, when it comes to keeping tabs on how broadly the virus is spreading, we could instead be using slightly less sensitive — though not poor quality — rapid tests, known as antigen tests, which typically can provide results in minutes to hours. Such tests should be used for screening people en masse in settings like nursing homes, essential workplaces, and communities that have limited testing resources, proposes a team at Duke University’s Margolis Center for Health Policy. Any positives that turn up could then be confirmed with a PCR test.

The goal is to avoid the long testing turnaround times that the country was plagued with this summer. PCR tests, while highly accurate, usually require at least a day or two to return results even under optimal conditions, and require more specialized equipment, labs and staff. This summer, when the majority of tests were being shoved into the PCR queue, turnaround times stretched out, with some people waiting more than two weeks for test results.

This is not just an annoyance for individuals. It’s a massive public health problem, because a test that takes more than two days to come back is pretty much useless.

“Patients don’t know what to do in those two weeks, and guess what, we can’t do our contact tracing, so we can’t fight the pandemic — all of that gums up the system.” said Shah, of Harris County. Such long turnaround times are “shameful. It makes no sense.”

Dr. Mark McClellan, one of the authors of the Duke paper, said the government must set aside funding to pay for antigen tests in at-risk populations, including low-income, minority and immigrant communities, and public schools and colleges.

The University of Illinois is requiring all faculty, staff and students to participate in screening testing twice a week, using a rapid saliva-based test. Not every college has the resources to perform these routine tests, but advocates for this kind of testing point to the university to show that it isn’t a fantasy.

“It is feasible,” said Carl Bergstrom, a computational biologist at the University of Washington. “It’s just a matter of will.”

McClellan and his co-authors estimate that about 14 million people are in high-risk settings that need regular screening testing, requiring an average of two tests per week. “There needs to be a lot more financial support to get that capacity up, something like Operation Warp Speed, with the government going in jointly with manufacturers,” he said.

What We Need to Do: Pick a Plan, and See It Through

For now, though, the federal government doesn’t appear to embrace this vision. Testing czar Giroir told reporters in a call on Aug. 13, “I’m really tired of hearing, by people who are not involved in the system, that we need millions of tests every day. … You don’t need this degree of testing. You need strategic testing combined with smart policies.”

Giroir explained that the administration’s focus was testing symptomatic patients as well as vulnerable populations, such as nursing home residents, coupled with policies including mask wearing, social distancing and hand washing. “That plan is being implemented and that plan is working,” he told reporters.

Some public health experts say that approach won’t be enough to curb the pandemic.

“Masks are a very powerful tool for virus control, and they’re not completely off the table, but a lot of our population has not been able to adhere to them because it’s become politicized,” said Dr. Michael Mina, assistant professor of epidemiology at the Harvard T.H. Chan School of Public Health.

And while social distancing is important, said Jha, he doesn’t think that alone will work in places where people are regularly congregating, like schools. “It’s not the real world,” he said. “Do we really think kids will never get close to each other?”

Mina argues for an audacious plan that calls for far more testing than the U.S. has been capable of to date. His testing strategy, particularly when it comes to how it approaches asymptomatics, seems directly at odds with Giroir’s.

Mina envisions tests so cheap ($1 apiece) and so widely available (over the counter) that every American can test themselves at least twice a week. The tests we’d use are paper strips that require only a saliva sample. They would certainly be less sensitive than PCR tests, but sensitive enough to catch people when their viral load is highest, which is exactly when they are most infectious.

The technology for a cheap, rapid antigen test certainly exists: Abbott Laboratories’ $5 test, authorized by the U.S. Food and Drug Administration last week, goes a long way to prove this point. But Abbott’s test is intended to be used on symptomatic patients, and needs to be performed by a doctor. Mina wants people to be able to test themselves.

Mina’s vision has gained broad support in recent weeks by numerous public health experts, but would need buy-in from the federal government, particularly the FDA, to become reality.

Many otherplans have been proposed, but at this point, more time has been spent talking about what we should be doing and debating the various options, rather than mustering the necessary regulatory, financial and political power to get any one of the plans fully executed.

“Choosing not to test those who are asymptomatic is like saying we won’t fight the fire until it reaches the second floor,” said Brian Castrucci, chief executive officer of health philanthropy the de Beaumont Foundation.

The pandemic has been raging across America for more than half a year. It’s past time we had a coherent national plan to put out the fire.