Coronavirus ‘test pooling’ could help, but not solve, the U.S. testing morass
Testing in batches could expand and speed up COVID-19 diagnosis while cutting costs. But don't count on a magic bullet.
The latest fix for what ails this country’s ability to promptly diagnose coronavirus infection is called “test pooling.”
It involves analyzing patients’ respiratory samples in a batch rather than individually.
So far, only two commercial lab companies, Quest and LabCorp, are authorized to do pooling, and the batches cannot have more than four or five patients’ specimens. Still, some experts are touting pooling as a panacea.
“For the price per test of a cup of coffee, governments can safely reopen the economy and halt ongoing COVID-19 transmission — all without building new labs and without new drugs or vaccines,” three Massachusetts Institute of Technology professors wrote in MIT Technology Review.
Pooling is such a hot topic that the Infectious Diseases Society of American held a virtual media briefing on Friday to discuss it and answer questions. The bottom line, said Melissa Miller, director of the clinical molecular lab at the University of North Carolina Medical Center, is that “this is not a magic bullet.”
Here are some advantages, downsides, and limits of test pooling.
More efficient — sometimes
More than 800,000 diagnostic tests per day are being done in the United States, according to the COVID Tracking Project. That still isn’t adequate to curb the pandemic, most experts say.
The surge of cases in hot spots, the lack of federal coordination of testing, and the dominance of a handful of commercial testing giants have combined to create testing backlogs. That, in turn, means people with no symptoms — even health-care workers at high risk of exposure — may not get results for a week or more. And that’s too late to guide decisions about quarantining and contact tracing.
By combining specimens for testing — which involves molecular analysis to detect fragments of the virus — results can be speeded up, and labs use less of the chemicals that have been in chronic short supply.
But this strategy only works if all the specimens are negative. If the pool tests positive, each sample must be retested individually to figure out who is infected, and that can delay results.
For pooling to work, a technician must carefully select the samples to merge based on certain variables. These include patient factors, such as having no symptoms or living where transmission is low, and the prevalence of the virus in the community. If 15% to 20% of tests come back positive in a given area, “that renders pooling impossible,” Romney M. Humphries, director of the microbiology laboratory at Vanderbilt University Medical Center, said during the media briefing.
The U.S. Food and Drug Administration says pooling is appropriate if the positive test rate in a given area is 10% or less, but Humphries believes it should be no more than 2%.
The three MIT professors, led by Ziad Obermeyer, propose that mining data with computers to develop predictive formulas can make the selection process easier. They are cofounders of a health-care consulting company that develops such machine-learning algorithms.
Alas, most labs don’t have the technology, including automation, to make pooling optimally efficient, Humphries said.
Less accurate — sometimes
Pooling effectively waters down the specimens, making it easier to miss a trace of the virus. That’s one reason the Food and Drug Administration allows only four or five specimens to be combined.
Pooling also requires the kind of molecular testing platforms that are found in high-complexity labs.
Recently, the FDA has authorized some molecular testing platforms that are small, portable, and give rapid results at the point of care — say, a doctor’s office or urgent care clinic. But these tend to sacrifice accuracy for on-the-spot results; the agency in May warned that the Abbott Point of Care test, used by the White House, may miss some infections.
“Test pooling is not an option with point-of-care platforms,” Humphries said.
Batch testing is proving valuable in academic research labs that are studying the prevalence and spread of the virus in various locations. Nonclinical research labs, however, don’t need FDA authorization to pool samples, and the results are not reported to public health authorities because the testing is not for diagnostic purposes.
“It’s a bit controversial,” said Miller.