HHS relaxed oversight of problematic Covid-19 tests despite being told of accuracy concerns

Early in the coronavirus pandemic, management at the North Hill Retirement Community — a sprawling high-end campus for older adults in suburban Boston — kept a wary eye on how the virus was sweeping through similar facilities with devastating, deadly consequences.

It took precautions to protect its residents, and when the order came down that by mid-July, all of its 400 staff members had to be tested for the virus every week, administrators scanned a list of approved tests provided by the Massachusetts health department and selected a vendor — a Boston-based DNA lab named Orig3n that had developed its own test for the novel coronavirus.

North Hill was soon to learn, not all Covid-19 tests are created equal. In a vast, confusing landscape populated by hundreds of different private companies and labs, some tests are proving problematic, especially when used to screen people who have no symptoms of Covid-19. A STAT investigation found that top officials at the U.S. Department of Health and Human Services ended already-minimal oversight of these so-called laboratory developed tests in mid-August, despite being informed that the tests were plagued with quality issues.

It was just two weeks earlier that North Hills administrators, after going months with only one positive test for Covid-19, had been informed that 19 staff members tested positive for the virus at the same time.

“It was really out of the blue,” said Ted Owen, North Hill president and CEO. “The numbers looked strange to us, that suddenly there would be this big thing.”

State public health investigators were called in to verify the test results, and meanwhile, Owens said, “There was a week of absolute pandemonium.” Staffers were sent home to nervous families, waiting for symptoms to emerge, while residents and their families waited to see if the virus had spread among the vulnerable, older population of North Hill. In the end, the problem was not a viral outbreak but a testing issue. All 19 of the results were false positives; none of the staffers actually had Covid-19.

North Hill’s false positives were not isolated. A state investigation in August found 383 false coronavirus positives in that period from Orig3n tests, which had been used by about 60 nursing homes in the state. The state health department suspended Orig3n’s testing, demanding the company address contamination and documentation issues. Orig3n told STAT it faults human error for its testing problems.

The Orig3n test is just one of dozens that hospitals, university researchers, and private labs rushed to make as the pandemic spread across the U.S. All these laboratory developed tests, known as LDTs, use the same basic PCR technology to identify genetic material from the coronavirus in patient swabs, but their performance can vary widely. They are being used across the country by states, private employers, sports leagues, and schools to try to manage the pandemic, yet they are something of a black box. Because they fall into a regulatory gray area, nobody outside the labs themselves knows how accurate they are — not even the health care providers who use them.

In February, as the coronavirus was starting to spread across the U.S., the FDA began requiring these labs to apply for an emergency use authorization, as it had in previous public health emergencies like the H1N1 flu pandemic in 2009. The bar was set low — to get tests on the market fast, labs were allowed to start using them without waiting for FDA sign-off — and the standards for approval were far less than would normally be required. Labs merely had to show reasonable accuracy detecting the virus on “contrived” test samples that weren’t obtained from actual patients.

Then this summer, the Department of Health and Human Services directed the FDA to stop requiring emergency authorization for this segment of the testing market. The Centers for Medicare and Medicaid Services retained oversight of the labs, but it ensures only that labs meet quality standards, not that specific tests work. The decision meant the FDA has no way to track problems with Covid-19 LDTs — including false-positive or false-negative results — or even which labs are marketing these tests.

An HHS spokesperson said the decision was made for purely legal reasons, based on its Office of General Counsel’s determination that the FDA does not have jurisdiction over lab-developed tests.

“This deregulatory action ensures compliance with law, is responsive to multiple Trump Administration Executive Orders and better prepares us for future pandemics while maintaining regulatory safeguards for quality and accuracy,” HHS chief of staff Brian Harrison said in a statement.

But STAT has learned that before the decision, HHS and its general counsel, Robert Charrow, were alerted to major concerns about many of the LDTs for detecting the coronavirus, which had been identified during the FDA’s review process.

Another FDA official, speaking on background, told STAT that data on problems with LDT accuracy and performance were sent by the FDA to HHS at the beginning of August, and were seen by Charrow before the decision to revoke FDA authority over LDTs, announced on Aug. 20. This official was on multiple phone calls with FDA colleagues and scientists in which the submission of data to HHS was discussed, including with people who had gathered and compiled the data and “who were expressing extreme frustration with HHS for, as they described it, suppressing the data.”

The data became public in a September article in the New England Journal of Medicine, in which two FDA officials outlined how the emergency use authorization process enables the agency to track the safety and problems with tests, but HHS’ earlier knowledge of the data has not been previously reported.

In the first months of the pandemic and testing rollouts, the FDA analyzed 125 EUA requests from laboratories, and found “design or validation problems” with 82, while “several have been denied authorization,” wrote FDA’s Jeffrey Shuren, director of the Center for Devices and Radiological Health, and Timothy Stenzel, director of the FDA Office of In Vitro Diagnostics and Radiological Health. In most cases, the EUA process allowed the agency to work with labs to correct problems, they noted.

“No test is 100 percent accurate and performance can vary within populations,” they wrote. “Covid-19 diagnostic tests may be less accurate in asymptomatic or low-risk populations and in persons who shed little virus or are early or late in the course of illness.”

In a later interview, Shuren said the FDA’s oversight role is critical in monitoring tests for safety and accuracy and keeping problematic tests off the market.

“The value of FDA oversight of tests is that we are able to review and rely on the science to better understand whether or not a product works and how well it works,” he said. “If we’re not reviewing something, we can’t do that.”

Responding to criticism of the HHS decision, Charrow, the general counsel, said in a statement, “It is false to say that this will lead to unregulated, low-quality COVID-19 tests. Every single COVID-19 test in the United States will still be regulated at the federal level. This simply removes an additional regulatory barrier … that at most other times is not required.”

Alberto Gutierrez, former director of the FDA Office of In Vitro Diagnostics and Radiological Health, who was with the agency for 25 years, said the current chaos in testing could have been prevented if the FDA and HHS were working together to track problems and safety and get the best tests to market.

“The problem with LDTs is there are a lot of different laboratories and there are no real standards because the tests are developed by the laboratories,” he said. “There’s no bar to meet. You have very good laboratories that are very thorough, and you have some that cut corners.”

In early October, three Democratic members of the U.S. House — Reps. Frank Pallone Jr. of New Jersey, the chairman of the Energy and Commerce Committee, Anna Eshoo of California, and Diana DeGette of Colorado — wrote to HHS Secretary Alex Azar, demanding answers about the agency’s decision to end FDA oversight and tracking of lab-developed tests. During the Covid-19 pandemic, they said, FDA jurisdiction over LDTs “is warranted” because as former FDA Commissioner Scott Gottlieb recently wrote, “in public health emergencies, diagnostic tests not only diagnose diseases that aren’t completely understood, but help guide analyses of disease progression and risks to public health.”

The testing company at the center of this summer’s debacle in Massachusetts — where it was also deployed to test people who are homeless — says it is working to improve its Covid-19 test, which remains off the market there and in other states where it was used, including North Carolina. “We’ve taken extraordinary steps to work through this challenging time and hope to be able to bring our capacity carefully back online to help address this pandemic,” said Orig3n CEO Robin Smith.

Orig3n was granted emergency authorization in April, before HHS dropped the requirement. Under an EUA, test makers are supposed to report to the FDA when things go wrong — including false results. But few do. A review of the agency’s public database for testing issues shows only 657 reports related to Covid-19 diagnostic tests as of Sept. 30, out of more than 100 million tests performed by that point. This means the vast majority of false Covid-19 results — including all of the Orig3n testing errors — are simply never reported.

The FDA acknowledges underreporting is rampant, limiting the database’s usefulness for spotting problematic tests. Especially for LDTs, it relies on essentially an honor system, expecting labs to report on themselves. Without better tracking and oversight from the federal government, some agency officials told STAT, they fear testing problems will continue as the pandemic worsens through the fall and winter, with patients and medical providers flying blind on which tests can be trusted.