After a bout with COVID-19 a few weeks ago, Jim Parkes has been feeling better lately, but he wasn’t sure he had fully shaken the virus and worried he might still be contagious.
So when Parkes, 74, a retired Mercer Island resident, heard that UW Medicine’s Virology Lab had started performing tests to check for coronavirus antibodies in blood, he jumped at the opportunity, he said.
“I thought this could provide some reassurance that I was OK,” he said.
But Parkes’ efforts to get the test were quickly doused.
A Swedish Medical Center clinic didn’t respond to his repeated calls asking for a test order, Parkes said, and at Virginia Mason, a health care worker told him the UW test wasn’t FDA approved, so it wasn’t available. When Parkes challenged that response, the clinic told him the test was offered only to certain patients who met specific criteria. Neither response was true, according to university lab officials.
Washington’s testing for the novel coronavirus, already plagued by a slow rollout, restricted availability and a lack of testing supplies, now faces a new problem. As antibody tests that can purportedly detect whether people had COVID-19 infections become more widely available, misinformation, confusion and suspicion over the tests’ reliability is limiting patients’ access.
Such throttled access extends even to the antibody test the UW has used to assess more than 6,000 blood samples and is considered among the better-performing tests in a dubious array of them. Despite general resistance to antibody testing among clinics in Washington and even Gov. Jay Inslee’s administration, the UW test is available to anyone.
“We just need a doctor’s order and a blood sample,” said Dr. Geoff Baird, interim chair of the UW’s laboratory medicine department.
But for Parkes, such as seemingly simple task wasn’t attainable. When Swedish finally returned his call this week, a nurse gave him yet another reason the UW test wasn’t available: The clinic didn’t know how order it.
“Each day, you get a different answer,” Parkes said. “You don’t know what the heck to believe. I finally just threw up my hands.”
“Wild West out there”
More than 1 million people in the U.S. so far have tested positive for the novel coronavirus, but experts largely agree that is a vast undercount due to slow and limited diagnostic testing in Washington and elsewhere. Some modelers estimate as many as 10 to 20 times more people have been infected.
Antibody tests, such as the one Parkes sought, differ from the diagnostic assays used to detect the virus’ genetic fingerprint via nasal swab, in that they look for proteins in blood serum that build immunity to infections. Theoretically, antibody tests can determine if people who became sick with COVID-19-like illnesses but were never tested — or those who never had any symptoms — at one time had the virus.
Until a COVID-19 vaccine is developed, public health officials, businesses and politicians may look to antibody testing to determine a level of “herd immunity” when assessing whether to ease the stay-home orders and allow individuals to safely return to work.
That’s why tests that routinely produce bad results could be dangerous, especially as questions and problems with the array of blood tests persist. More than 70 firms have developed COVID serology tests, but experts say their quality ranges wildly and includes some that simply don’t work.
“It’s the Wild West out there, and I say caveat emptor,” Dr. Jeff Duchin, health officer for Public Health – Seattle & King County, said during a recent webinar sponsored by Life Science Washington. “The UW test is excellent, but there’s a bunch of garbage out there as well.“
Virology lab director Dr. Keith Jerome said recently UW scientists reviewed “a number of tests” before settling on the test from Abbott Laboratories, a global biomedical firm based in the Chicago area.
The company and UW separately validated the test’s performance with similar results — finding it was more than 99.6 % specific (the rate of correctly identifying specimens without antibodies) and 100% sensitive (the rate of correctly identifying antibodies in positive specimens) about two weeks after the onset of symptoms.
“It is very, very specific,” Jerome said. “If we tell you that you have the antibodies, there’s an extremely high likelihood that those antibodies actually came from COVID rather than something else.”
More than 6,000 blood samples have been tested since April 23, with about 5,000 of them from Idaho showing a 1.8% COVID prevalence rate. The first few dozen samples tested from the Washington region returned a 4% to 10% prevalence rate, but the samples tested “are not at all representative” of the state population, and don’t yet give an accurate picture for the disease’s spread here, Baird said.
With enough testing, Baird said, scientists plan to eventually design studies to “get very good estimates of how many people in Washington were infected.”
“We regret any confusion”
After her older brother, Eric Braunberger, an Issaquah nursing home resident, came down with pneumonia that eventually killed him, Lisa Roberts got sick and had flu-like symptoms at the end of January. But Roberts, 68, of Renton, doesn’t know if she or her brother were sickened by COVID-19 because Swedish never tested him, and her symptoms cleared by the time testing ramped up.
To find out, Roberts recently called UW Medicine, got instructions for how to order the test and sent the paperwork to her doctor at a local Kaiser Permanente Washington clinic. But “they just flat-out can’t or won’t” order the test, she said. Roberts eventually found a private clinic in Renton that would.
Other people have experienced similar problems. One Swedish physician repeatedly denied that her clinic could even order it, according to written exchanges with a patient reviewed by The Seattle Times. The doctor contended the UW itself had informed Swedish last Friday not to send it any samples, before reluctantly offering to arrange an antibody test through Swedish’s preferred reference lab, LabCorp.
“We are not aware of any communication to Swedish not to send us tests,” Baird said.
A Swedish spokeswoman said in emails this week the hospital system hasn’t directed its physicians to deny requests to order the UW test, or to favor LabCorp. Still, Swedish isn’t routinely ordering antibody tests, and is only doing so “if needed — working with our partners at LabCorp,” she said.
LabCorp and Quest Diagnostics, two of the nation’s largest industrial labs, only recently offered broad antibody testing. Quest launched a physician-ordered test last week in which patients can get through clinic visits for $55, and a home-testing service for $119. LabCorp’s test was made available this week through Walgreens locations in at least 11 states, as well as medical clinics.
By comparison, the test offered by UW costs $42 to perform, Baird said. Such tests typically are covered by insurance.
Overlake Medical Center offers only the LabCorp test to patients. Other health care systems in Washington, including Virginia Mason, Kaiser Permanente Washington and PeaceHealth, said they are not providing antibody tests, generally citing reliability issues.
“The COVID-19 situation is evolving and information is changing rapidly, which could explain why an individual might get different explanations about the serology test,” Virginia Mason spokesman Gale Robinette said. “We regret any confusion a patient has felt.”
“Just not true”
Misinformation about the UW test isn’t isolated to local clinics. A spokesperson for Washington’s coronavirus Joint Information Center, when asked this week about the state’s position on broad antibody testing, didn’t differentiate among available tests and broadly characterized them as “not clinically usable.”
“The tests currently out are meant for offices or other noncertified lab locations, which makes interpreting their accuracy claims more challenging,” a JIC spokesperson told The Times in an email.
“That’s just not true,” countered Baird, after reviewing the state’s email. “The test we are using are provided by a very large biomedical company and they are very much intended for clinical laboratories like our own.”
Because the UW isn’t manufacturing or selling the tests, just performing them in-lab, Food and Drug Administration approval isn’t needed, Baird added. Such Laboratory Developed Tests, or LDTs, are regulated by the Centers for Medicare & Medicaid Services, under which more than 70 antibody tests have been developed. Nonetheless, the FDA has conducted preliminary reviews of some of them, granting emergency authorizations for at least eight, including the Abbott test.
A group of three dozen conservative-leaning Washington state lawmakers have pushed Inslee to call for more widespread antibody testing and surveying as a way to help decide whether and when social-distancing restrictions can be eased.
“This could be a valuable data point in the long run — especially if [cases] come up again in the fall,” said Sen. Lynda Wilson, R-Vancouver, who is leading the effort. But so far, Wilson has gotten “radio silence” from the governor, she said.
Inslee’s office said it’s keeping on eye on antibody testing, but his primary focus remains on massively expanding diagnostic testing and contact tracing to better identify and isolate those infected.
“Antibody tests are not useful for diagnosing active infections, as even a high-quality test will typically be negative in the early days of infection, and a positive test may indicate historical rather than active infection,” said Reed Schuler, one of Inslee’s senior advisers.
Antibodies to two other coronaviruses — Severe Acute Respiratory Syndrome and Middle East Respiratory Syndrome — last between one and two years, but scientists haven’t yet determined how effective they are in protecting against COVID-19, said Dr. Nahid Bhadelia, a Boston University medical school infectious-disease physician.
“What we do know is that survivors of this current disease, COVID-19, do develop antibodies that seem to be directly targeted, effectively, at blocking the virus,” she said during a press briefing last week. “Is that enough? How high does that level need to be?”
UW scientists have realistic expectations about what exactly the test can and cannot do. The test can evaluate whether a patient has developed antibodies to the virus, Baird said, “but what we don’t know yet is whether the antibody confers immunity or protection.”
Correction: An earlier version of this story did not include information that blood samples tested so far by the UW included more than 5,000 specimens from Idaho that found a 1.8% COVID prevalence rate.