In the midst of the 2016 Zika outbreak, a Centers for Disease Control and Prevention (CDC) virus expert blew the whistle. He alleged the CDC was promoting a test that was far less reliable than another in use at public health labs across the U.S. He further charged that the agency had concealed this information from the broader public health community.
But he directed his most fundamental critique at what he saw as an increasingly bureaucratic and centralized approach at the CDC that hamstrung testing for quickly spreading pathogens.
“A CDC scientist warned years ago that the federal government’s virus testing efforts would be too slow during a rapidly moving outbreak.”
“Time will tell how effective it is and what unforeseen problems arise” with the CDC’s testing approach, the whistleblower, Robert Lanciotti, wrote in 2016 in an official response to a government investigation of his concerns. Calling the agency’s approach “rigid,” he wrote that an approach where the public health laboratory community is reliant on the CDC to manufacture and then ship out test kits was “a departure from the long standing method of test dissemination by the CDC.”
“This lack of flexibility is more than merely a weakness of this approach; it is in fact a fatal flaw,” Lanciotti wrote, arguing that “rapid essential adaptability” was “required during public health emergencies.”
His critique proved prescient. Dr. Anthony Fauci, the head of the National Institute of Allergy and Infectious Diseases, told Congress in March that it was “a failing” that coronavirus testing was not far more widespread in the U.S. much earlier.
Lanciotti, who left the CDC at the end of 2018, after 29 years with the agency, told the Project On Government Oversight (POGO) that the CDC did not address his concerns.
“They doubled down instead,” Lanciotti said. “Coronavirus is proof that their approach has failed.”
“Coronavirus is proof that their approach has failed.”Robert Lanciotti, CDC whistleblower
Unlike Zika, which is primarily transmitted through mosquito bites, the coronavirus can be transmitted through the air and by people who do not show symptoms. The relatively greater speed, stealth, and ease of the coronavirus’s spread has magnified the consequences of inefficiencies in the public health response, including how long it took to ramp up testing. The ongoing pandemic highlights the balance the government must strike between rapid response and flexibility and ensuring safety and effectiveness.
POGO’s review of the federal response to Zika and the 2009 H1N1 influenza pandemic shows there were warning signs that during a serious outbreak the federal system for ramping up testing might move too slowly. Yet during those emergencies an approach that allowed state public health, hospital, and academic labs to use tests they made themselves mitigated issues with the rollout of CDC test kits and the speed of Food and Drug Administration (FDA) authorization of other tests. Without that flexibility during the coronavirus outbreak—in conjunction with an unforeseen snafu with the CDC’s test kits—the shortcomings of the system were laid bare.
Last year, well before the new coronavirus emerged, the federal government launched a task force to speed up testing in a public health emergency, citing inadequate testing during the Zika outbreak. But there’s no indication that the task force—involving the CDC and the FDA—heeded Lanciotti’s warnings that centralizing testing at the CDC during the initial phase of an outbreak could be disastrous.
Neither the CDC nor the FDA responded to POGO’s detailed queries.
CDC Director Robert Redfield has appeared to downplay the effect of the delay in widespread testing, telling reporters on May 29, “we were never blind” to detecting the spread of coronavirus. He cited a recent CDC study that stated that “overall disease incidence before February 28 was too low to be detected” by hospital emergency departments. Yet the study’s authors acknowledged their findings were limited because widespread testing was not “immediately available after discovery of the virus.”
Some experts argue that the lack of testing had the effect of delaying government responses—with deadly consequences. Had policymakers put social distancing measures in place one week earlier in March, at least 35,000 lives could have been saved in the U.S., according to Columbia University researchers.
“Single Point of Failure”
The federal government’s coronavirus testing efforts failed in February for two main reasons, and the scope of the problems extends beyond the CDC.
First, state public health laboratories received test kits from the CDC that had been contaminated during manufacturing at a CDC facility and did not produce reliable results. (The CDC could still conduct some tests at its headquarters in Atlanta.)
Then, as the country largely lost weeks’ worth of testing and the virus continued to spread, the FDA did not authorize the use of other tests. Until February 29, when the FDA issued a new policy significantly loosening regulation of tests, the CDC’s flawed test was the only one authorized in the U.S.
Some experts have described the reliance on the missing-in-action CDC-manufactured test as the pandemic response’s “single point of failure.” In essence, the government placed all its chips on one test and kept losing for weeks.
“The biggest challenge here was not at the FDA,” Josh Sharfstein, deputy FDA commissioner from 2009 to 2011, told POGO. “It’s hard for the FDA to judge, independent of the CDC and HHS [Department of Health and Human Services], how the public health situation is evolving.”
During past outbreaks, input from the CDC appears to have played a role in how tightly the FDA regulated testing, according to Lanciotti, a 2009 federal advisory committee review, and a 2011 Association of Public Health Laboratories report.
“It’s hard to imagine the FDA contacting public health labs and commercial manufacturers to call for a Manhattan Project for testing while the White House is saying coronavirus isn’t a big deal,” Sharfstein said, referring to the World War II-era project to develop the atomic bomb.
A large-scale testing effort also likely would have required the White House’s approval to seek new appropriations from Congress—and reports indicate that in early February White House officials balked at doing so. The White House requested appropriations later that month, and more funding for testing by state labs came in early March.
And meanwhile, Sharfstein said, he believed there were few applications in February from state labs to the FDA to authorize other coronavirus tests for use because the lab community “may have been waiting on the CDC” to fix and provide its test kits for much of that month.
Indeed, the Association of Public Health Laboratories sent a letter on February 24 to the FDA stating, “while we appreciate the many efforts underway at CDC” to send test kits to state labs, “this has proven challenging and we find ourselves in a situation that requires a quicker local response.” The association requested that the FDA make it easier for labs to use tests other than the CDC’s.
At the end of the month, the FDA issued its new policy, giving labs the green light to use tests they made themselves.
States’ dependency on a CDC-manufactured test for almost the entirety of February did not have to happen, Lanciotti told POGO.
Lanciotti said “there are a lot of parallels” between what went wrong with coronavirus testing earlier this year and the concerns he raised in 2016.
During both the Zika outbreak and the coronavirus outbreak, more complex CDC tests were associated with accuracy problems.
However, the CDC was more permissive of alternative tests during Zika, even as its favored test was found by some experts to miss a high percentage of positive cases.
For years, Lanciotti led a CDC lab in Fort Collins, Colorado, that specialized in viruses spread by mosquitoes, such as Zika. He played a role in responding to West Nile virus, which spread in New York state in 1999.
So when Zika made its way to the U.S., first to Puerto Rico in December 2015, then to Florida in July 2016, his lab was ready. Lanciotti’s lab and at least 20 state public health labs used the test design that Lanciotti and his colleagues created and made their own tests, none of which were authorized by the FDA.
In the fall of 2015, as the virus spread in South America—appearing to cause birth defects, such as neurological problems, in infants born to mothers who had been infected while pregnant—a CDC lab in Puerto Rico had developed a different Zika test. That test also detected two other mosquito-borne illnesses, dengue fever and the chikungunya virus. In the spring of 2016, the CDC’s Emergency Operations Center, which leads the agency’s response to health crises, chose Puerto Rico’s test to submit to the FDA for authorization. The FDA authorized it in March 2016, and the CDC began manufacturing and shipping out test kits.
Some labs continued using tests based on the design from Lanciotti’s lab. Public health labs in Tampa and Jacksonville, Florida, used it because CDC kits “were not initially available,” according to an account of Florida’s Zika testing. Those Florida labs did not switch to the CDC test kit until December 2016.
When comparing the two tests in April 2016, Lanciotti’s CDC lab and a private company found that the FDA-authorized test produced false negatives more often. An email from a representative of that company said the test missed up to a third of Zika cases, calling the results “disturbing.”
But the CDC did not share this information with state public health labs. Soon after, Lanciotti blew the whistle, sparking a Department of Health and Human Services investigation led by Stephan Monroe, the CDC’s head of lab science and safety.
While the CDC lab in Puerto Rico reached inconclusive results when comparing the two tests, in August 2016, the CDC changed the more complex test to require a larger testing sample. After that, the more complex test’s accuracy became comparable to that of the one developed by Lanciotti’s lab, according to the Department of Health and Human Services and a peer-reviewed study that Lanciotti later coauthored along with 14 other scientists.
In a September 2016 report, Monroe ultimately sided with the CDC’s decisions to favor the more complex test and withhold information from state labs, and cited the changes in August. The report said sharing information about the varying reliability of the two tests would have created “broad uncertainty across public health laboratories about which diagnostic test they should use.”
Another reason the CDC-led investigative report defended the more complex test was that it was tightly controlled by the agency. Labs could only use test ingredients “provided by CDC, ensuring CDC oversight of the manufacture and quality control of the materials,” the report stated.
Similarly, until the FDA made its policy changes for coronavirus testing at the end of February 2020, labs could not simply take the CDC’s test design and make and administer their own tests. The CDC had posted its design online, but the agency told labs in February that they could not use other tests without FDA authorization, even if their tests were based on the CDC’s design.
Without FDA-authorized alternatives, labs were dependent on the CDC to manufacture and distribute test kits. On February 5, two weeks after the first known U.S. case of COVID-19, the disease caused by the new coronavirus, was confirmed, the agency began shipping out kits with its three-part test to state public health laboratories. Labs soon found that the test was unreliable. CDC officials later confirmed that contamination of the test in a CDC facility caused the problems and the FDA found the “CDC did not manufacture its test consistent with its own protocol.”
Yet the part of the test kit that was failing wasn’t needed.
On February 28, the CDC and the FDA gave state public health labs the go-ahead to jettison the third part of the test. “The third component, which was the cause of the inconclusive results, can be excluded from testing,” a senior CDC official said at a news briefing.
The third part of the test was intended to distinguish the new coronavirus from other coronaviruses, such as one that causes severe acute respiratory syndrome, or SARS. But SARS isn’t spreading around the world, was last detected in the U.S. in 2003, and is significantly genetically distinct from the virus that causes COVID-19.
“Either the CDC didn’t know it was a crisis or they should have pulled the plug on that [the third part of the test] faster,” Paul Keim, a genetics expert at Northern Arizona University, told the Washington Post.
Standing next to President Donald Trump at CDC headquarters in Atlanta on March 6, the CDC’s Monroe, who had investigated Lanciotti’s concerns years earlier, spoke to reporters. Glossing over the lost month of testing due to the CDC’s problems, Monroe said, “We quickly went from designing the test, and the materials were made in these laboratories. The quality control is done here.” He also said that the CDC’s test was created to distinguish coronavirus from “other kinds of viruses,” a reference to its third component.
The immediate breakdowns with the CDC’s coronavirus tests are “human mistakes,” Lanciotti told POGO, but “the bigger mistake is the structure of testing that puts CDC at the center.”
“While it looks good on paper, in reality, in responding to an epidemic, it’s much too slow, too cumbersome,” he said.
During the initial phases of an outbreak when a virus appears to be spreading rapidly or has the potential to do so, basing the entire testing strategy on centrally making a test kit and shipping it out, as the CDC did with its coronavirus test this year, is not the best approach, Lanciotti told POGO. He said it’s not the nimblest way for the nation’s public health system to leverage existing government and other labs to initially scale up testing.
“For rapid dissemination, it’s better to send a protocol,” Lanciotti said, which he and others also refer to as a “recipe,” along with details on where to buy the ingredients needed. The CDC and logistics experts in the federal government should also shore up the supply chains for the materials needed to make the tests, he said. (This approach does not necessarily preclude the CDC from also making and shipping out its own test kits.)
Lanciotti told POGO that disseminating the recipe and materials to public health labs was how the CDC got a test out for West Nile virus.
The West Nile test, like the CDC’s coronavirus test, the two Zika tests described above, and many others, used a set of techniques referred to by the acronym RT-PCR, or reverse-transcription polymerase chain reaction.
“PCRs are the easiest tests—little can go wrong except for contamination,” Lanciotti said, provided that the labs have adequate experience, the right materials and instruments, and meet certain standards.
Lanciotti said that during his time at the CDC, he and his colleagues regularly assessed whether state public health laboratories could successfully use PCR techniques to test for viruses. Lanciotti said that nearly 100% of the time, the labs were successful.
“I always saw CDC as a gap-filler,” to play a critical role early during an outbreak in partnership with public health labs, before commercial manufacturers developed tests authorized by the FDA and rolled out on a larger scale, Lanciotti said.
He’s not the only one at the CDC who has argued for this approach.
In a 2011 Association of Public Health Laboratories report on the rollout of testing during the H1N1 flu pandemic of 2009, Dr. Daniel Jernigan, who currently leads the CDC’s influenza division, made similar remarks.
“Before, the federal government would have to make a lot of tests, and we’d send them out, and people would use them—that was the full process,” Jernigan said in the H1N1 report. But being able to distribute the recipe for a test to labs makes for “a much more robust approach for people able to ramp up testing for a novel pathogen.”
During the H1N1 outbreak, then-FDA Deputy Commissioner Sharfstein granted an emergency-use authorization for a CDC-developed test. This authority can only be used during a declared public health emergency. While it lowers the bar for temporary FDA approval for medical devices and drugs relevant during the emergency compared to the agency’s normal approval process, an emergency declaration creates higher hurdles for labs to use tests they make—unless the FDA exercises its enforcement discretion.
“As H1N1 shows, public health labs’ ‘home brews’ can be critical in a crisis.”Association of Public Health Laboratories
When there is not a declared emergency, labs can generally use what are known as “laboratory-developed tests” without FDA authorization, although the labs still must meet standards enforced by the Centers for Medicare and Medicaid Services.
Sometimes called “home brew” tests, these can be developed entirely by one lab, or a lab can base its test on a recipe developed by the CDC or others. A home brew test can be a minor modification of a test that has been authorized by the FDA.
Though the FDA quickly authorized the CDC-developed test, a Department of Health and Human Services report on H1N1 said most of the FDA’s emergency use-authorizations for additional tests came “well after the peak demand for testing” meaning testing capacity was limited when it was needed most.
Because of that limited capacity, the CDC apparently asked the FDA to use its discretion to allow labs to continue using tests that had not been granted emergency-use authorization. “When the FDA was considering putting H1N1 tests under more control and scrutiny, the [CDC] task force picked up the phone again” to tell the FDA to hold off, according to the Association of Public Health Laboratories report.
Even though some labs developed tests with information “quickly released by the CDC,” a federal review of the H1N1 response found that the broader unavailability of tests “led to frustration within the clinical community” and “made infection control more difficult.”
This February, Jernigan, who was regularly communicating with the state public lab community, reportedly advocated for the FDA to loosen its policy in order to make it easier for labs to conduct their own coronavirus tests, according to the Washington Post.
Some companies have exploited the FDA’s track record of generally not regulating laboratory-developed tests. A 2015 FDA report laid out 20 case studies where laboratory-developed tests, some marketed by companies, harmed or could have harmed patients.
“In some cases, due to false-positive tests, patients were told they have conditions they do not really have, causing unnecessary distress and resulting in unneeded treatment,” the report stated. “In other cases, the LDTs [laboratory-developed tests] were prone to false-negative results, in which patients’ life-threatening diseases went undetected.”
The Silicon Valley startup Theranos, founded by Elizabeth Holmes, infamously promised a rapid test for a variety of diseases based on a pinprick of blood. But Theranos could not overcome key technical and scientific barriers, despite raising more than $700 million from investors.
Holmes told “investors that FDA approval was not necessary because she believed that Theranos’ tests were laboratory developed tests,” according to a 2018 Securities and Exchange Commission (SEC) complaint against Holmes and the company, alleging they misled investors. Despite being warned by Walgreens—with which Theranos had a business agreement—that FDA approval seemed necessary, Holmes and Theranos marketed the test to “retail patients” without approval, states the complaint. Theranos and Holmes settled with the SEC without admitting or denying the agency’s claims.
In a public health emergency, there are heightened societal consequences of allowing shoddy tests to proliferate. People who don’t believe they are contagious may be less likely to take measures to prevent the spread of a pathogen. And, as is true even in normal times, unreliable testing can mean that people who need treatment don’t receive it.
But time is of the essence during a fast-moving outbreak and detecting cases quickly can be crucial to halting widespread infection. The Association of Public Health Laboratories’ 2011 report stated that “as H1N1 shows, public health labs’ ‘home brews’ can be critical in a crisis. Can they stand up to regulatory scrutiny—or can a response wait for a regulatory review?”
“Trouble and Confusion”
During the Zika outbreak, those working with the CDC’s Emergency Operations Center appeared to prioritize getting the FDA’s sign-off over ensuring the state labs were using the most accurate tests they could.
In an email exchange obtained by POGO, one of Lanciotti’s colleagues wrote him and others at CDC in April 2016 that state labs believe “they are running the ‘best’ test because it is FDA approved” but “ironically, the issuance of the EUA [emergency use authorization] has prevented them” from running what appeared to be a substantially more accurate test. Another CDC colleague wrote, “The simplest resolution might be to convey this information to the states and let them decide.”
Believing that the state public health laboratories needed to know, Lanciotti emailed state labs, telling them that his CDC lab in Fort Collins was still using its simpler test, and briefly noted that he had information indicating that it was more reliable in detecting Zika.
Not long after, the CDC demoted Lanciotti from being in charge of his lab. He then filed a whistleblower retaliation complaint with an independent government watchdog called the Office of Special Counsel, which intervened to have the demotion reversed. (Disclosure: The author is a former Office of Special Counsel employee.)
“It shouldn’t matter what test kit we’re using or what procedure as long as we can get the right answer for patients.”Testimony of a hospital pathologist
An official detailed to work with the CDC’s Emergency Operations Center, Ann Powers, told Lanciotti in an email that his message to state labs created “more trouble and confusion.” She then emailed the state labs without mentioning the accuracy differences between the tests.
“We acknowledge that some laboratories may have created in-house laboratory developed tests (LDTs) … some employing sequences published by CDC scientists. Laboratories should understand that these LDTs are not covered by the FDA” emergency-use authorizations, Powers wrote.
But she left an opening, writing that these home brew tests could still be used without FDA authorization if the labs took certain steps. The Department of Health and Human Services’ investigative report following Lanciotti’s whistleblower disclosure said the Emergency Operations Center’s decision to recommend the FDA-authorized CDC test kits “while not prohibiting the use of other validated laboratory diagnostic tests … was a reasonable and appropriate course of action.”
Yet the report noted that the FDA took a starker position: “FDA does not believe its policy of enforcement discretion is appropriate for laboratory developed tests offered in public health emergencies, including the public health challenge involving the Zika virus.”
“Enforcement discretion is not guaranteed,” the report stated, and it is the FDA, not the CDC, that has the official power to let labs use tests that it had not authorized.
During a September 2016 congressional hearing, then-Senator Orrin Hatch (R-UT) said test developers from his state encountered burdensome and shifting requirements from the FDA as they sought an emergency-use authorization for their Zika test.
“It shouldn’t matter what test kit we’re using or what procedure as long as we can get the right answer for patients,” a hospital pathologist testified in response.
“It’s All Enforcement Discretion”
The FDA’s position that it had to authorize tests during a public health emergency seemed to stick through early 2020.
In February, a University of Washington lab tried to obtain FDA approval for its test, but the FDA wanted proof that the lab’s test could also distinguish the new coronavirus from other coronaviruses that are not known to be circulating in the U.S.: SARS and Middle East respiratory syndrome.
Obtaining a SARS sample to test would have taken more than a month, University of Washington professor Alexander Greninger told ProPublica. In a February 14 email to his colleagues, Greninger wrote, “the most pernicious effect of the current regulatory environment is that it kneecaps our ability for preparedness.”
In contrast to the CDC’s more permissive message during the Zika outbreak, the CDC presented FDA authorization as the only option. On February 18, the CDC told state public health laboratories that “All tests … including laboratory developed tests (LDTs), must be reviewed and cleared or authorized by the FDA for emergency use, or they cannot be used for diagnostic testing.”
Six days later, the Association of Public Health Laboratories urged the FDA to use its “enforcement discretion” to allow state labs to create and use “a laboratory developed test (LDT) for the detection of” the coronavirus. Allowing state labs to develop their own tests would offer “a more expeditious route” than the FDA’s emergency-use authorization process, the association wrote.
The FDA reversed course a few days later, issuing its policy allowing labs to immediately start using their own tests even without authorization, provided they submit an application to the FDA seeking approval. A government document dated February 27 obtained by the Washington Poststates the plan “for achieving more rapid testing capacity” involved giving laboratories latitude to “design their own tests and make or order the individual components.”
“It’s all enforcement discretion,” Sharfstein, the former FDA deputy commissioner, told POGO. “The LDT [laboratory-developed test] policy is enforcement discretion.”
Referring to the FDA’s actions to allow the use of tests that had not yet been authorized, “they did what they had to do under very difficult circumstances,” said Sharfstein.
“There is no normal order of business in a crisis like this.”Josh Sharfstein, former deputy FDA commissioner
But it can be a challenge to strike the right balance between increasing the availability of testing during the crisis and promptly ensuring that tests produce reliable results.
A rapid test produced by the company Abbott Laboratories and used by the White House to test Trump and others who come in close contact with him or the vice president has come under fire for appearing to miss up to 48% of positive cases, according to academic research that has not yet been peer reviewed. Abbott has defended its test by saying other studies show it is as accurate as other tests. The FDA, which granted an emergency-use authorization for the test in late March, has said it is studying the test.
In addition, Sharfstein said the FDA “opened the floodgates to antibody tests,” also known as serological tests, which are generally less accurate. “They are now moving to tighten up there.”
Like the elusive test promoted by Theranos, antibody tests use samples of blood. By detecting the presence of antibodies in blood, which show that a person’s immune system has fought off an infection, they can show whether someone has been infected. These tests tend to be used retrospectively rather than for people showing symptoms or with other reasons to believe they have recently been exposed to a pathogen.
The FDA increased its scrutiny of antibody tests in early May. “We unfortunately see unscrupulous actors marketing fraudulent test kits and using the pandemic as an opportunity to take advantage of Americans’ anxiety,” two FDA officials wrote.
Hindsight is 2020
“There is no normal order of business in a crisis like this,” Sharfstein said. But he said the FDA could have had a more permissive policy for coronavirus testing earlier if the Department of Health and Human Services had conveyed the urgency to do so. He noted that would have contradicted the administration’s position at the time.
Brett Giroir, the Health and Human Services official who was put in charge of testing in mid-March, told the Washington Post the FDA should have moved sooner. “In retrospect, it might have been useful earlier, right?” he said.
Now a Catholic deacon, Lanciotti said one of his biggest disappointments with his former agency is that during the Zika outbreak the CDC was not as transparent with the public health community as it should have been and withheld information on the reliability of its favored test versus the simpler test.
“In retrospect, it might have been useful earlier, right?”Brett Giroir, federal official overseeing coronavirus testing
The Government Accountability Office, the investigative arm of Congress, agreed with Lanciotti. “The lack of access to data on test performance prevented users from making informed decisions about which test to adopt or recommend during the outbreak,” according to the watchdog office’s 2017 report. The watchdog recommended the CDC make test data available even for tests the FDA has not authorized. The CDC disagreed, saying it is “impossible” to reliably compare tests that haven’t received FDA authorization.
With the COVID-19 death count in the U.S. now topping 100,000 and the pandemic far from over, policymakers must reevaluate many elements of the government’s public health response. Among them: the government’s approach to testing.
“It seems to make sense until you’re on the front lines and it falls apart,” Lanciotti said.