COVID testing: More isn't necessarily better, 'it’s all about how smart they are testing'
The nation’s slow, fumbling response to the pandemic spawned a new mandate: Get tests to the public as quickly as possible.
Lack of testing for COVID-19 in February and March left the nation unprepared for a virus that raced ahead of the ability to detect it. The Trump administration sought to accelerate testing by relaxing bureaucratic red tape and spending billions to bolster labs, drive-thru clinics, retail testing sites and rapid point-of-care testing at doctor's offices.
With public and private labs now performing more than 400,000 tests most days – a figure the U.S. Department of Health and Human Services projects will more than double by late summer – some question whether accelerated testing comes at a cost. A strategy built on the quantity of tests does not account for nuances and limitations, lacks coordination and wastes limited testing supplies.
“I get concerned that in the rush to get testing out, we’re compromising quality,” said Dr. Eric Blank, chief program officer of the Association of Public Health Laboratories. “You’ve got to have quality in order to make informed decisions.”
Tests inform governors and public health officials on when it's safe to resume daily activities such as going to restaurants, social gatherings or sporting events. School boards must decide when and how students can return to classrooms. Employers want to safely reopen offices.
President Donald Trump has repeatedly touted the U.S. as the world’s testing leader, setting a goal of 40 to 50 million tests each month by September. However, the Trump administration wants states to develop their own testing strategies to slow the COVID-19 pandemic, which has so far infected nearly 1.9 million U.S. residents and killed more than 108,000.
Test accuracy varies and can give individuals who falsely test negative for the virus an unwarranted sense of security. Labs are operating around the clock to keep up with demand from the expectation that any person, at any time, can get a test. And with three types of tests from dozens of test-kit makers, doctors and patients are confused by when and what type of test they might need.
Dr. Robert Redfield, director of the Centers for Disease Control and Prevention, acknowledged Thursday that better testing involves more than increasing the number of tests.
"I don’t want to get into the number of tests because I don’t think that’s the real issue," Redfield told a Congressional committee Thursday. "It’s how testing is used and what’s the consequence."
Others say it’s not the numbers of tests that are important. Rather, it’s how tests are deployed and coordinated to gather meaningful information about the pandemic.
Dr. Michael Osterholm, director of the Center for Infectious Disease Research and Policy at the University of Minnesota, said the nation needs a smarter testing system. CIDRAP, a public health and infectious disease research group, has asked HHS Secretary Alex Azar to appoint a panel of national experts to address testing objectives.
Meanwhile, the Trump administration has stressed states need to implement their own testing strategies. On April 27, the administration released a testing blueprint, and reiterated states' testing responsibilities in a May 24 report to Congress called the COVID-19 Strategic Testing Plan.
The administration set a goal of states testing 2% of their population, but the report showed state testing rates varied widely.
Osterholm said focusing on the quantity of tests might not yield the best results. For example, a drive-thru clinic might test several hundred individuals with no signs of the virus, while people with symptoms can’t access testing.
“Far too many states that have focused on, 'Just how many can I get tested today?' ” said Osterholm, an infectious-disease doctor.
The key is to develop a system that gets the right test to the right person at the right time, he said.
“If two states have the same population, one state might test 5,000 per day, the other 10,000 people per day, and the state at 5,000 per day might be doing a much better job,” he said. “It’s all about how smart they are testing.”
'We are not getting answers'
Infectious disease doctors say testing is critical because the virus will continue to spread until an effective vaccine is available. Without a vaccine, the virus might spread beyond 5% now infected to 60% to 70% of the population needed to reach herd immunity, Osterholm said.
"So with all the pain, suffering, death and economic disruption that’s occurred, we’ve got a long ways to go yet," he said.
Until then, public health labs that provide surveillance testing will play a vital role in tracking the virus and slowing its spread.
In February, the CDC sought to establish a tracking system in six cities to track whether the virus had spread beyond initial travelers returning from China. The effort stalled, leaving a patchwork system tracking spread in communities.
Private labs are testing. Large retailers such as Walmart and CVS are making it convenient. And doctors and individuals get the results.
But there’s no centralized effort to collect information in uniform way, said Dr. Farzad Mostashari, CEO of Aledade Inc. and a former senior official with HHS.
“The most obvious failure is we don’t have any consistent testing of a consistent population that we can track over time in different places,” said Mostashari, also a former assistant commissioner of the New York City Department of Public Health.
Such uniform data could answer key questions: What is the average time to diagnose a patient from the time symptoms appear? How can contact tracers get timely information needed to slow the spread?
“We are doing mass testing at Walmart and we’re celebrating how great it is,” said Mostashari. “Even though we are doing magnitudes more tests, we are not getting meaning. We are not getting answers.”
Accurate, public health data would allow decision-makers to react quickly as states and cities reopen. Contact tracers can inform infected individuals to isolate and quickly notify their contacts to monitor for symptoms.
“What you really want to know is who has it now and is capable of giving to someone else today,” said Dr. Robert Wachter, chair of the University of California, San Francisco Department of Medicine.
UCSF is testing thousands of patients, even those without symptoms, to get a snapshot of the pathogen's community spread. The academic hospital is conducting surveillance testing in two San Francisco communities: Mission District, a neighborhood with a large Latino community and a higher prevalence of COVID-19, and Bayview Hunters-Point, which has a large African American population.
In the absence of a federal testing strategy, the hospital is trying to build data and learn more about how the virus is spreading. Such information will help discern when the virus is heating up and allow local leaders to adapt.
“That is kind of the super power of this virus – you might have it and be capable of transmitting it and you feel fine,” Wachter said.
'Each test has different flaws'
There are three types of coronavirus tests – molecular, antigen and serology – each with varying degrees of accuracy.
The molecular and antigen tests detect whether a person has an active case of COVID-19, the disease caused by the SARS CoV-2 virus.The serology tests detect whether a person has been previously infected with the virus and has developed antibodies to fend it off.
Given that there are three types of tests with varying degrees of accuracy, consumers and even some doctors might be confused, experts say.
“People look at a test as a test, like turning a light switch on and off,” Osterholm said. “It’s not that simple.”
One measure of test accuracy is sensitivity: how often a test correctly identifies a person has the virus that causes COVID-19. Another measure is specificity: whether the test accurately shows when the person does not have the virus.
Doctors need to evaluate characteristics of each test and apply reasoning when interpreting results, Wachter said. For example, a doctor should consider a patient's clinical symptoms and community trends of the virus.
“Each test has different flaws,” said Wachter.
Molecular PCR tests, which require a mucus sample from the nose or throat, typically have low levels of false positives. So when a molecular test for the virus is positive, it’s a safe assumption the test is right. However, when a person tests negative for the virus, that does not mean they are disease-free.
Timing is critical. A person might not immediately develop symptoms, so they don’t get tested. There is a limited window of five to seven days when the test will detect a person has coronavirus. Or an inadequate sample might fail to pick up enough genetic material to identify the virus.
“The test doesn’t get it wrong that often when the test is positive,” Wachter said. “Whereas it does get it wrong reasonably often, when the result is negative.”
Antibody tests are blood tests that detect whether someone was previously infected. Public health officials raised concerns about test accuracy after the Food and Drug Administration allowed test makers to validate their own data before marketing to consumers. Following complaints, the FDA moved to tighten oversight of these tests.
By September, as many as 25 million tests will be rapid "point of care" tests at clinics, doctor's offices, nursing homes and urgent care centers that deliver immediate results, according to the COVID-19 Strategic Testing Plan.
An example of this type of test is the Abbott ID Now test that can deliver results within 15 minutes. The FDA issued an alert about "early data that suggest potential inaccurate results" from the toaster-sized device following a small study from New York University.
But Abbott questioned the NYU study and released its own data from an interim study that found the test identified the virus 94.7% of the time. Test results depend on factors such as patient selection, specimen collection, handling, storage and transport, Abbott said.
Wachter said the accuracy of the Abbott ID Now test, used by the White House, is an example of decision employers and doctors must consider when evaluating a test. The White House uses the tests to decide who can come into contact with Trump, who choose not to wear a mask.
It’s important to get correct test results so infected individuals can self-isolate and not pass the virus to others. And public health officials can contact those who were potentially exposed to the infected individual.
Still, the push to get rapid testing to the public might compromise quality, Blank said.
"If you want something rapid, if you want something quick, if you want something accessible, you are going to have to give something up," he said. "But the question I have is how much quality can you afford to give up?
Public health labs stressed
Blank, of APHL, said public health laboratories are struggling to keep up with testing demands of the pandemic.
State and local public health laboratories are accustomed to running full tilt for two to three weeks to respond to local or regional outbreaks. With coronavirus, public health labs have been running extra shifts seven days a week since February.
“We have heard consistently that they are holding up – but barely,” Blank said.
Large commercial labs are accustomed to running round-the-clock shifts with testing platforms that can quickly process samples. Public health labs don’t have the staffing and equipment to sustain those levels.
Redfield said Thursday he's "personally saddened" only a handful of state labs have "the capability to do what needs to be done." He praised New York's Wadsworth Center lab for developing rapid testing on a "high-throughput platform," not the "slow platform we have for flu."
The CDC sent $10.25 billion to states to invest in public health response. Each state must submit detailed testing plans and benchmarks to the CDC – the first report was due May 31, and a second report is due June 15.
State labs also need sustained funding to hire trained labs workers to replenish an aging and stressed workforce.
“They’ve never come out and said we are at the breaking point,” Blank said. “I never expect them to say that. But I have been a lab director and I suspect they are close to the breaking point.”
Follow the reporter on Twitter @kalltucker or email him at firstname.lastname@example.org