NEW YORK – The National Institutes of Health Rapid Acceleration of Diagnostics initiative has launched a new study aimed at determining how at-home rapid antigen testing for SARS-CoV-2 compares with lab-based molecular testing for detecting the virus.
The effort, called the Test Us At Home initiative and led by the University of Massachusetts Chan Medical School, aims to enroll several thousand subjects around the country, each of whom will regularly perform at-home SARS-CoV-2 tests and submit samples for PCR testing, allowing researchers to compare results of the two methods.
The PCR testing will be done by Quest Diagnostics out of its Marlborough, Massachusetts laboratory, while the at-home testing part of the study will use the Abbott BinaxNow, Quidel QuickVue, and BD Veritor over-the-counter tests.
The study launched at the end of October and will run from between one and several months depending on how long it takes to enroll enough subjects and register enough positive cases to compare the accuracy of the two testing methods, said William Heetderks, a RADx Tech program advisor at the National Institute of Biomedical Imaging and Bioengineering, or NIBIB.
Heetderks said the study was recruiting subjects across the country, focusing on areas with high or rising rates of COVID-19 cases.
The study builds on similar work sponsored by NIH and led by researchers at the University of Illinois at Urbana-Champaign, which likewise compared the performance of rapid antigen and molecular SARS-CoV-2 tests. That study looked at samples collected from 43 individuals at the University of Illinois, where on-campus students and staff are required to undergo saliva-based SARS-CoV-2 molecular testing every two to four days. The researchers ran saliva- and swab-based PCR testing as well as swab-based rapid antigen testing using Quidel's Sofia SARS Antigen FIA test and the Sofia 2 device.
They found that while PCR testing had higher sensitivity than antigen testing (roughly 98 percent for both saliva- and swab-based PCR tests versus 80 percent for antigen testing), more frequent testing with antigen assays compensated for this difference. Testing using antigen tests every three days raised antigen assay sensitivity to better than 98 percent, the equivalent sensitivity of weekly testing using the molecular assays.
While these findings provided insight into how useful rapid antigen tests could be for catching COVID-19 cases, they didn't directly address the question of how effective OTC rapid antigen tests are, given that the Quidel test used was a point-of-care assay that relies on a reader. The Test Us At Home study will look specifically at OTC tests, helping researchers account for factors like user sampling and reading errors and potential performance differences between instrument-based and instrument-free rapid antigen tests.
"This is intended to absolutely duplicate the typical user experience," Heetderks said.
He noted that there is a help line available to study participants, but that "we have talked quite a bit about what we can and can't do on that help line in terms of making the study, as much as possible, look like what the naïve user would do."
Some studies have indicated that rapid antigen test performance varies significantly depending on who is doing the sampling. For instance, data collected by UK researchers as part of SARS-CoV-2 rapid antigen test evaluations found that rapid lateral flow antigen assays from Pasadena, California-based test vendor Innova Medical Group showed 77 percent sensitivity in validation studies (compared to PCR testing) when used by trained staff, but only 58 percent sensitivity when used by self-trained members of the general public.
Bruce Tromberg, director of the NIBIB and leader of the RADx Tech program, said the Test Us At Home program study would add to data that was being collected by the Say Yes! COVID Test project, an initiative run by the NIH and the US Centers for Disease Control and Prevention that has been studying the usage patterns and effectiveness of SARS-CoV-2 home tests.
Launched in April, the effort initially distributed free at-home tests (primarily Quidel's QuickVue test) to residents in Pitt County, North Carolina, Chattanooga, Tennessee, and Ann Arbor and Ypsilanti, Michigan, and has since expanded to provide free tests to Fulton County, Georgia, O'ahu Hawaii, Louisville, Kentucky, and Indianapolis.
Tromberg said that in some of these communities — the Michigan sites in particular — participants were offered the opportunity to enroll in follow-up studies, including those that assessed the performance of the rapid antigen tests compared to PCR. He noted, though, that the study was not by its design powered statistically to answer this question, whereas the Test Us At Home study will be.
"There are multiple questions in the whole [issue] of how does testing reduce the chain of transmission of the virus," Tromberg said. "The first is, can people actually use them the right way and get as good a result from self-sampling as you would get with a professional.
"And then there is sort of the analytical head-to-head comparison for a given prevalence and pretest probability. How do these tests really work with respect to false positives and false negatives, and what is the impact of that?" he added. "It really looks at a combination of these things that I think are really important."
While the Test Us At Home study aims to provide data on the effectiveness of home tests versus lab-based PCR, the Say Yes! COVID Test project has focused more on understanding home test demand and usage patterns, with a particular emphasis on underserved communities and areas at high risk of COVID-19 spread due to factors like low vaccination rates or high numbers of essential workers.
The program made 2 million rapid antigen tests available at the North Carolina and Tennessee sites, using broadcast media as well as community groups and leaders, including churches, to make people aware of the availability of testing and to help with test distribution. It made 1 million rapid antigen tests available to residents at each of the Georgia, Hawaii, Kentucky, and Indianapolis sites.
Between 8 percent and 14 percent of the population at each of the locations requested tests, said Rachael Fleurence, senior advisor, immediate office of the director at NIH, noting that in O'ahu, where 14 percent of the population ordered tests, demand outstripped supply, which artificially lowered the percent of residents participating.
At the end of October, researchers from the project published data in Health Affairs from a market research survey of residents of the North Carolina and Tennessee sites. Among the findings were that around three-quarters of residents at both sites were aware that home COVID-19 tests were available and around half were aware that they were available for free through the initiative. Roughly half of the respondents at both sites also said that their primary concern about the tests was that they might be less accurate than lab-based tests.
While studies have generally shown that OTC rapid antigen tests are less sensitive for detecting SARS-CoV-2 than lab-based molecular tests, proponents of home testing have argued that more frequent testing with the OTC assays can compensate for their lower sensitivity — an assertion that studies like the U of Illinois study have borne out.
The Say Yes! COVID Test project indicates, however, that few people are actually using home tests in a serial manner. According to the survey data published in Health Affairs, only 11 percent of the North Carolina participants and 12 percent of the Tennessee participants used the tests according to the serial use instructions in the test kit.
"The way most people use the test is if they are going out and they want to make sure they are not infectious or if they think they have been exposed," Fleurence said. "What we're seeing is that people are using them more on an as-needed basis versus following a kind of regular screening pattern."
She noted that while in this iteration of the study, the instructions call for regular testing two to three times a week, "in our next iteration we will be slipping into something more flexible where people will just use them as they feel they need to."
Much of the work comparing the effectiveness of antigen SARS-Cov-2 testing to PCR has used a serial antigen testing model, but Fleurence said she and her colleagues believe "the tests are still useful in asymptomatic patients" even when used at single points in time as opposed to serially.
"I think generally we are comfortable with the rapid antigen test [used at single points in time] being pretty sensitive to detecting infectiousness, whether that individual is symptomatic or asymptomatic," she said.
She added that agencies like the CDC "were not there yet" in terms of sharing this perspective on rapid antigen tests. Noting she was not speaking for NIH or another government agency, Fleurence said, "I think maybe the jury is still out, but there is a school of thought that says that comparing sensitivity of rapid antigen directly to PCR is not appropriate because PCR also picks up that long tail where you have the dead virus in your systems and you are testing positive but are no longer infectious."
She acknowledged that data from the University of Illinois study indicated that rapid antigen testing would miss positive COVID-19 cases early in their infectious period that molecular testing would catch, but said that that "is a very small window" of time.
"The viral load does increase pretty quickly, even through that first day of infection," she said.