News

How the CDC Botched Basic Science in Its Attempt to Make a Coronavirus Test

“It actually boggles the mind.”
A clinical support technician extracts viruses from swab samples so that the genetic structure of a virus can be analyzed and identified in the coronavirus testing laboratory at Glasgow Royal Infirmary, on February 19, 2020 in Glasgow, Scotland. (Photo by

The Centers for Disease Control and Prevention sent out its first batch of COVID-19 diagnostic test kits on February 6, equipping public health authorities around the country with what should have been the capability to rapidly test up to 50,000 Americans for the disease. That same day, two planes carrying around 350 evacuees from Wuhan, the Chinese city at the epicenter of the novel coronavirus outbreak, landed in California.

Advertisement

One sick passenger, who was quarantined at a military base in San Diego, initially had their test results come back negative. But less than 24 hours later, after re-testing the sample at agency headquarters in Atlanta, the CDC realized there had been a mistake, a technical snafu — the person, who had been kept in isolation as a precaution, was positive after all.

At the time, it seemed like a relatively minor mistake. More than a thousand people had already died from COVID-19 in China, but there were only about a dozen cases in the U.S. and none of them had been fatal. Dr. Nancy Messonnier, director of CDC’s National Center for Immunization and Respiratory Diseases, explained the situation to reporters by calling it a “mishap.”

“The mishap was unfortunate,” Messonnier said on February 12, “but we have corrected this from happening again in the future by adding additional quality control.”

But by February 8, it was obvious that something had gone seriously wrong with the CDC’s testing kits. Around 50 labs had received the kits, and 36 reported having issues. The problem appeared to be with chemicals known as reagents, which are used to process samples from nasal swabs for lab analysis. Unable to trust results from the kits it had just distributed, the CDC was forced to continue doing all tests in Atlanta until new reagents could be manufactured.

READ: The U.S. has a classified stockpile of ventilators and masks, but it won’t save us from coronavirus.

Advertisement

The CDC’s botched testing rollout, which contributed to the undetected spread of the virus, has proven to be extremely costly. There are now over 5,000 confirmed COVID-19 cases in the United States and 92 deaths. Millions of people are under orders to shelter-in-place in a desperate bid to contain the outbreak. The Trump administration has since taken steps to scale-up U.S. testing capabilities, but it remains extremely difficult even for people showing symptoms to get tested.

The testing fiasco has been the subject of intense media scrutiny in recent weeks, but the CDC still hasn’t explained exactly what went wrong and why. But with the information that is publicly available, scientists have identified a technical design flaw that could have potentially led to erroneous results — just like the American evacuee from Wuhan.

“I thought I could learn something, but it was the exact opposite. I was catching mistakes.”

“This process has not gone as smoothly as we would have liked, but there is currently no backlog for testing at CDC and since Friday, February 28, 2020, more and more state labs are coming online, increasing our public health system’s ability to detect and respond to cases,” a CDC spokesperson said in a statement to VICE News. “If a state or local public health lab partner has been/is able to validate the CDC diagnostic lab test, then we are comfortable with the accuracy of the results.”

Advertisement

The Department of Health and Human Services did not respond to an inquiry from VICE News, and the CDC did not address specific technical questions about what went wrong. Independent experts who spoke out about the situation cautioned that it’s impossible to say for certain what happened without more details. But there is strong evidence that suggests the CDC bungled an initial step in the testing design process that most scientists consider to be rudimentary. The task simply requires punching a sequence of letters into one of several widely available computer programs, which are designed to catch errors and ensure optimum efficiency.

“It actually boggles the mind,” one expert wrote on Twitter, describing the magnitude of the CDC’s mistake.

Tomer Altman, a Stanford-trained bioinformaticist who runs his own consulting company in San Francisco, said he first noticed the problem when the CDC released guidance and resources for professionals who are involved in studying the coronavirus or testing samples.

“I was just curious,” Altman said. “I thought I could learn something, but it was the exact opposite. I was catching mistakes.”

The issue has to do with the CDC’s choice of “primers,” which are one of the reagents used in the testing process. When Altman input the CDC’s novel coronavirus primers into a computer program, he immediately realized something was wrong.

READ: Coronavirus is forcing AA meetings to close, and people are worried about relapse.

Advertisement

“I just put the primers as input into this program and within seconds it tells you there’s problems with it,” Altman said. “It’s freely available. It’s not hard to use at all. It’s one of those things you’d do as a basic quality control check. It seems like they hand-designed these primers and it seems like they didn’t use computer tools to check it.”

For reasons that remain unclear, the CDC opted to design its own COVID-19 test with new primers despite the fact that the World Health Organization had already approved a test developed by German scientists. The tests were in widespread use elsewhere without any of the technical problems seen in the US. South Korea, which detected its first positive case one day prior to the U.S., was conducting over 18,000 tests a day by March 3. As of March 12, the U.S. had conducted fewer than 14,000 tests total.

The COVID-19 tests use a technology called PCR (short for polymerase chain reaction), which has been around since the mid-1980s. After a sample is collected from a nasal swab, reagents are used to extract RNA, which contains the virus. Another step, known as reverse transcription, converts the RNA to DNA. The next step is when primers enter the equation.

“The CDC released primers that were, simply put, terribly designed.”

One nasal swab sample yields a multitude of DNA strands, which are essentially lines of genetic code. Primers detect specific DNA sequences within that code that are unique to the virus. There are multiple sets of primers that can be used in this process, but some work better than others. After using the primers, a scientist can process the results manually in the lab or, more commonly, use a machine to get the final result.

Advertisement

When Benjamin tenOever, a microbiology professor at Icahn School of Medicine at Mount Sinai who specializes in the study of viruses, used the CDC primers for an experiment in his lab last week, he found that they were “picking up all kinds of nonsense.” The results should have provided a clear signal about whether or not the virus was present, but instead they were cluttered by background noise, a flaw that could, in a worst case scenario, produce false positives even in samples with no virus present.

“The CDC released primers that were, simply put, terribly designed,” tenOever told VICE News. “I think this is a disgrace given how easy this task is and this error seems to be a good representation of how poorly the CDC has handled this crisis.”

While he cautioned that there’s no definitive proof that the primer flaw is the root cause of the CDC’s malfunctioning test kits, tenOever said CDC’s mistake is glaring nonetheless. “I will give them the benefit of the doubt and say the test is fine, but the first step is adding the probability of getting false positives,” he said.

Selecting the primers to detect a virus, tenOever added, is a basic task typically assigned to young students who are just beginning their training in his lab. It’s as if the CDC made a rookie mistake. “It could have been so much easier and so much better than it was,” he said. “The CDC should have vetted its product a lot better than it did before publishing it.”

Advertisement

A CDC spokesperson said that after state and local health labs began reporting problems with the test kits on February 8, the agency “began working on a solution and identifying the source of the problem, which may include assay design or contamination.”

The initial unreliability of the CDC test kits contributed to critical delays that forced all samples to be sent to Atlanta for testing. The issues have since been corrected, and testing is now being done locally and at private lab facilities, but the early delays came at a critical juncture. Without effective and widespread testing, it was — and still is — impossible to know for sure how quickly the virus is spreading and where. Based on the available data, one expert in Seattle estimates that there are around 20,000 cases in the United States already, almost four times the number officially confirmed by testing.

“Lives are on the line. We can’t be making mistakes that other outside experts are saying, ‘Hey, these are basic mistakes’.”

Aside from technical glitches, the COVID-19 testing has been plagued by bureaucratic red tape. One scientist at the University of Washington, Dr. Helen Chu, had the infrastructure in place to conduct widespread testing in Seattle in the first week February, but was told to stand down by federal regulators. When her lab began testing samples anyway on February 27, they detected a positive case almost immediately, preventing an infected teenager walking into his high school. Despite the success, Chu told the New York Times that she was subsequently ordered to shut down her testing operation.

Advertisement

The first case of community transmission was found in California on February 26, but the Food and Drug Administration, which works closely with the CDC, didn’t authorize state and local labs to do initial testing on their own until three days later. It took until March 3 for the CDC to lift all restrictions, with the FDA commissioner claiming the U.S. would have capacity to test a million people by the end of that week. High demand has since contributed to a shortage of the reagents needed for testing, making it almost impossible to deliver on that promise.

Elected officials are already calling for a formal investigation into what exactly went wrong at the CDC and with the Trump administration’s response to the outbreak writ large. Washington Sen. Patty Murray, a Democrat, has gone so far as to push for a 9/11 Commission-style probe to get to the bottom of what happened.

“We have needed to work quickly since day one to slow the spread of this virus and every moment the federal government has struggled to provide adequate testing capacity is a moment we’ve been falling behind,” Murray said in a statement to VICE News. “We’ve heard administration officials give different estimates and timelines for testing and say anyone can get tested — but we know from people on the ground that just isn’t accurate, it’s not even what the experts are currently recommending.”

Murray sent a letter to Vice President Mike Pence, the Trump administration official in charge of overseeing the coronavirus response, on March 5, demanding answers about the testing rollout. One Senate staffer familiar with the situation told VICE News that, as of last week, “We still don't have any clarity about what went wrong and why [the CDC] didn’t use the WHO test.”

President Trump himself has said he bears “no responsibility” for the debacle, and seemed unaware last week that his administration disbanded a special National Security Council unit on global health security. That unit, which was shut down in 2018, was created in the wake of the 2014 Ebola outbreak specifically to deal with the US response to global pandemics.

The Trump administration has since formed a new task force of officials with the specific goal of rapidly ramping up COVID-19 testing capabilities. The federal government was expected to begin shipping materials on Monday that would allow for creation of mobile or drive-thru testing centers. Brett Giroir, a top deputy to health secretary Alex Azar, said 1.9 million tests would soon be available to as many as 2,000 laboratories nationwide.

Altman, the San Francisco bioinformaticist, told VICE News he hasn’t heard from anyone at the CDC since posting a detailed analysis of the flawed primers on his website on March 3. He’s still hoping for some clarity about how the CDC screwed up so badly.

“There's no transparency about the process,” he said. “Lives are on the line. We can’t be making mistakes that other outside experts are saying, ‘Hey, these are basic mistakes’.”

Cover: A clinical support technician extracts viruses from swab samples so that the genetic structure of a virus can be analyzed and identified in the coronavirus testing laboratory at Glasgow Royal Infirmary, on February 19, 2020 in Glasgow, Scotland. (Photo by Jane Barlow - WPA Pool/Getty Images)