MARTIN BERGSMA / SHUTTERSTOCK / GETTY / THE ATLANTIC

The US Has Tested Fewer than 5,000 People for the Coronavirus — and That's a Big Problem

By this point in its outbreak, South Korea had tested more than 100,000 people. The lack of data hurts U.S. government, corporate, and personal attempts to make decisions.

Nearly two weeks after the new coronavirus was first found to be spreading among Americans, the United States remains dangerously limited in its capacity to test people for the illness, an ongoing investigation from The Atlantic has found.

After surveying of local data from across the country, we can only verify that 4,384 people have been tested for the coronavirus nationwide, as of today at 4 p.m. eastern. These data are as comprehensive a compilation of official statistics as currently possible.

The lack of testing means that it is almost impossible to know how many Americans are infected with the coronavirus and suffering from COVID-19, the disease it causes. While our analysis has tracked state and local announcements that more than 570 people in 36 states are infected, experts say that number is almost certainly too small to reflect the full extent of the disease’s spread in the U.S. Not enough Americans have been tested for officials to know how many people are ill, they say.

When researchers have used statistical and genetic techniques to estimate the true size of the outbreak, they have concluded that thousands of Americans may have already been infected by the beginning of the month. Health officials have attributed 26 deaths to COVID-19 in the United States, as of today.

The sluggish rollout of the tests has become a debilitating weakness in America’s response to the spread of the coronavirus. By this point in its outbreak, South Korea had tested more than 100,000 people for the disease, and it was testing roughly 15,000 people every day. The United Kingdom, where three people have died of COVID-19, has already tested more than 24,900 people.

The Atlantic reached its new estimate through an ongoing collaboration with the data scientist Jeffrey Hammerbacher and a team of volunteers recruited for their experience with data collection, and after consulting data published by all 50 states and the District of Columbia. States vary widely in their reporting standards. All provide positive case reports. But many do not provide negative or pending case reports, which provide crucial context for both the progression of the virus and the government response to it.

Related: The Strongest Evidence Yet That America Is Botching Coronavirus Testing

Related: Defense Firms Limit Travel, Adapt As Coronavirus Spreads

Related: Iran Has Far More Coronavirus Cases Than It Is Letting On

Our effort is necessary because the Centers for Disease Control and Prevention is not regularly providing data on the full scope of American testing. On its website, the federal agency now provides a number (1,707 as of Sunday) that reflects only the number of people tested at the CDC’s laboratory, even though state and private laboratories provide the bulk of testing. (The CDC did not immediately respond to a request for comment.)

When the CDC has provided data, it has been slow and incomplete.

On Saturday, Stephen Hahn, the commissioner of the Food and Drug Administration, told reporters that 5,861 specimens—not people—had been tested for the coronavirus by the end of the week. As a rule of thumb, it takes about two specimens to deliver results for a single patient, which would make this equivalent to about 2,900 people tested through Saturday.

Last week, The Atlantic reported that it could only verify that 1,895 people had been tested for the coronavirus as of Friday morning.

Testing capacity still varies enormously across the country. Many states, including some of the country’s most populous, are not reporting how many tests they have conducted overall. Texas, which now has 24 positive cases, has not posted on its website how many people it has tested overall. A spokesman for the state said it had tested 150 people as of last week, but “with private labs coming online now, I don’t think we’re going to have a definitive number for the entire state going forward.” Nevada has not reported any new data at all on its health-department website since March 3.

Massachusetts, which has 41 presumptive cases, has not released its total number of people tested. Neither has Pennsylvania, which has 10 presumptive cases. Last week, a Pennsylvania official told us that the state could test only a dozen or so people a day, suggesting that it has a high rate of positives.

On Friday, California also stopped reporting how many tests it has conducted, switching to releasing only the number of positive cases.  The California Department of Public Health told us that the state had tested 778 people as of Saturday, and that the state has 114 positive cases. It now has 15 labs doing tests across the state.

North Carolina, which has two positive cases, and Indiana, which has two, have also never said how many overall tests they have conducted.

LabCorp and Quest, two companies that run routine medical tests for doctors’ offices, have both announced that they can now test samples for COVID-19. The two companies can test a combined 2,500 patients a day, according to a tally assembled by Gottlieb, the former FDA commissioner, and published by the American Enterprise Institute.

Altogether, the country can test a maximum of 7,840 people a day, according to Gottlieb’s preliminary tally. His count is another example of the kinds of data tabulation that a federal agency might usually take responsibility for.

The testing situation is so bad that Marc Lipsitch, an epidemiology professor at Harvard, says that health officials and journalists should stop reporting the number of positive cases in the United States as “new cases.” Instead, he wrote by email, “they should refer to them as ‘newly discovered cases,’ in order to remove the impression that the number of cases reported has any bearing on the actual number.”

The ponderous rollout of tests—and the stringent criteria that the CDC has imposed on them—has hamstrung doctors and injected anxiety into the lives of ordinary Americans. Are their symptoms pneumonia, the flu, or something worse?

“I have no clue if we could have already or could be now spreading this to others,” a 38-year-old woman who lives near Austin, Texas, who asked not to be identified for privacy reasons, told us.

After returning from Western Europe in late January, the woman and her husband came down with a mysterious illness, which sent them in and out of week-long fevers. She and her husband would wake up coughing in the middle of the night, their ribs aching so badly that they needed to vomit. She has tested negative for the flu, twice, and also tested negative for strep. She has been diagnosed with pneumonia.

On her trip, she had frequently been in large, international crowds, where she could easily have been exposed to the coronavirus. But despite having all the symptoms, she has not been tested for it. When she called Austin’s public-health department to ask for guidance, she was told that unless she was hospitalized or had traveled to China, she could not be tested for COVID-19.

“The woman who I talked to said, ‘There aren’t any cases here [in Travis County],’” she told us. “And I said, ‘There hasn’t been any testing, so how do you know?’”

Without a firm answer about whether she has the virus, she has agonized over how to act responsibly. When is she overreacting? When is she being reckless? She and her husband have stayed home since they became ill, but their son and daughter, both younger than 5, attended school until her daughter ran a fever last week.

“There’s no guidelines out there, even at the urgent care today,” she said. She now plans to keep both kids at home for the next two weeks.

But that’s only one of many arenas where there is currently no firm guidance for people who think they may have the virus, but who cannot get tested for it.

“Am I supposed to tell my team [at work]? Am I supposed to tell my kids’ school? Am I supposed to tell everyone I interacted with for the last four weeks?” she asked. “I don’t want to start a crisis, because I don’t know if I actually have this thing.”

Doctors have expressed similar frustration in getting patients tested. “The Georgia Department of Public Health has basically thrown up their hands when it comes to testing patients who do not require hospitalization,” Josh Hargraves, an emergency-room doctor in Georgia, told us. “On Friday we were told, ‘If the patient doesn’t have a travel history and doesn’t need to be admitted to the hospital, don’t bother calling; we’re not going to test.’” By Saturday evening, when Hargraves saw four prospective coronavirus patients, he managed to get one of them tested, but only after filling out onerous and unusual paperwork.

“We’re still restricting usage and asking thoughtful, knowledgeable medical professionals to jump through hoops to get a test they know a patient needs,” Hargraves said.

The outbreak is not at the same stage in every state. If public-health officials can quickly increase testing, it might be possible to have a much more comprehensive view before community transmission worsens.

We know the virus is here and spreading in many places. Restrictive testing policies—especially ones focused on travel outside the United States—clearly don’t make sense anymore. There are sick people in this country whose doctors think they need testing and who still cannot be tested. Every day that this epidemic continues without adequate testing, the country’s ability to slow the outbreak will deteriorate.

NEXT STORY: Ep. 63: War in space