U.S. blueprint highlights sequential COVID-19 antibody testing as route to minimizing false positives
BioCentury is providing this story for free given the urgent need for information about the COVID-19 crisis. For more analysis, sign up for our daily email.
In its COVID-19 Testing Blueprint unveiled Monday, the Trump administration made its latest push to improve the accuracy of serological tests on the U.S. market, this time by doubling them up. But banking on sequential testing to overcome the limitations of individual assays requires keeping paired tests as independent as possible, and an at least two-fold jump in testing capacity.
The document also divvied up responsibilities across different levels of government and the private sector.
Jointly issued by the White House, CDC and FDA, the blueprint’s core principles include timely and accurate diagnostic testing of every symptomatic patient, sentinel monitoring of at-risk populations, and containment of infections via contact tracing and isolation. It also said the U.S. should prioritize testing resources to hot spots, and take advantage of new rapid, point-of-care tests (see “Crisis Spurs Rapid Technologies”).
A focus of the blueprint was illustrating how administering two sequential COVID-19 serological tests could improve detection of patients previously infected with the virus over single tests administered on their own.
The double testing strategy, highlighted by FDA’s Timothy Stenzel in a virtual meeting last week, comes as reports indicate the first wave of serological tests flooding U.S. and European markets are failing to meet performance criteria. Stenzel is director of FDA’s Office of In Vitro Diagnostics and Radiological Health (see “World Awaits Validation of COVID-19 Serology”).
FDA has given companies the option market COVID-19 serological tests without regulatory review. But recent policy announcements by CMS and HHS, along with a test validation initiative led by NIH’s National Cancer Institute, indicate the Administration is seeking to improve the quality of COVID-19 antibody tests in the U.S., in part by subjecting them to greater scrutiny (see “New Policies Complicate Regulatory Calculus”).
The challenges involved in making tests that capture SARS-CoV-2 antibody responses across diverse populations, but don’t cross-react with responses to other viruses, have been made more acute by the COVID-19 crisis’ urgency, which has pushed companies to optimize and validate tests as quickly as possible.
FDA has justified its hands-off regulatory approach by arguing that the tests are not used to make clinical diagnoses. But since serological tests are expected to guide decisions such as where to allocate resources and who can more safely return to work, reports of their inaccuracy have raised serious concerns, particularly about the potential for false positives that give individuals a misplaced sense of security (see “Wild West of Antibody Tests”).
A test’s positive predictive value (PPV) -- the probability that a positive result is real -- is a function of the prevalence of true positives in a population. While there is still a lot of uncertainty about the prevalence of COVID-19 exposure, a commonly cited estimate for the U.S. is 5%. According to the U.S. blueprint, at that prevalence a test with 95% sensitivity and 95% specificity would only have a PPV of 50%, rendering its accuracy a crapshoot.
However, the blueprint’s calculations showed that sequentially administering two tests with that same sensitivity and specificity increases PPV across a wide range of prevalence values; at a 5% prevalence, the PPV of two positive tests is 95%.
Figure: Sequential COVID-19 serological tests
The figure shows how the positive predictive value (PPV) -- the probability that a positive result is real -- for single and sequential tests with 95% sensitivity and 95% specificity, under different assumptions about the prevalence of COVID-19 exposure in the population.
The results are based on the logic that each test has its own probability of generating a false result; multiplying those two probabilities together gives an overall lower probability of inaccuracy. However, the degree to which this holds true is a function of how independent the tests are from each other -- for example, whether they detect different antigens from SARS-CoV-2.
Even if tests are not completely independent, double testing will likely provide some measure of increased accuracy.
But a key question is whether it will be feasible to apply the strategy broadly in light of strains on testing resources, such as raw materials and laboratory capacity.
Who does what
The Trump administration said the federal government’s responsibilities are to provide strategic direction and technical assistance on testing, surveillance and contact tracing, and work with states to allocate testing supplies and capacity in accordance with need.
Federal agencies will also provide expedited regulatory authorizations for tests and equipment; publish guidances that specify prioritization algorithms and protocols for administering diagnostic tests; accelerate R&D for innovative tests by partnering with the private sector; and act as the test supplier of last resort.
It will be up to state, local and tribal governments to develop and implement the testing plans, sentinel systems and rapid response programs called for in the White House’s guidelines, as well as maximize the use of available testing providers, and identify and overcome barriers to efficient testing.
The blueprint calls upon the private sector and professional societies to develop new technologies, seek Emergency Use Authorizations (EUAs) for them when appropriate, accelerate production of tests and associated materials, share data from clinical trials, and expand testing partnerships with local, state and tribal governments.
-Robin Sawka contributed to this report.
Further analysis of the coronavirus crisis can be found at https://www.biocentury.com/coronavirus