Non-Prescription Home Molecular SARS-CoV-2 Test

Against the RT-PCR comparator, the home test showed a positive percent agreement of about 90% and a negative percent agreement of about 97%. The apparent false positives were attributed to subjects misinterpreting the result on an older version of the companion app, which was later modified and shown to reduce this error. One of the few apparent false negatives was negative on a second highly sensitive EUA RT-PCR assay.

In the near-cutoff lay-user study, PPA was about 94% and NPA was 100%. These results were added to the authorized labeling, together with app updates, to satisfy conditions of the original authorization.

FDA issued the Emergency Use Authorization for non-prescription home use and later reissued it. The EUA has since been revoked and is no longer in effect.

The roughly 90% PPA exceeds the 80% PPA that FDA's home-use template recommends for OTC symptomatic use, and its lower bound is above 70%. The NPA point estimate of about 97% sits just under the template's 98% recommendation; FDA authorized the test on its review of the complete submission, including the investigation of the discordant positives described here. Those discordant positives were traced to app-assisted misreading, and the app was changed in response, which shows how software and labeling count toward clinical specificity in a home test. With only about 30 comparator positives and about 80 negatives, a single result moves either estimate by several points, which is why the uncertainty is wide.

The near-cutoff study answers a different question from the clinical study. It shows whether lay users can detect low-positive samples near the assay's limit, which a small all-comer study may contain few of. Its roughly 94% PPA rests on about 20 contrived positives, so it supports the clinical result but does not replace it. The discrepant RT-PCR result is reported only as a footnote, consistent with FDA's practice of not using discrepant testing to recalculate PPA or NPA.

The authorization covered non-prescription home use. It included self-collected anterior nasal swabs from individuals aged 14 or older and adult-collected swabs from individuals aged 2 or older suspected of COVID-19, plus serial testing of people without symptoms (twice over three days, at least 24 and no more than 48 hours apart). Because the test would be run entirely by untrained users at home, the clinical evidence had to come from intended users performing every step themselves: sample collection, amplification, and reading the result through the companion app.

Performance had to be established against a high-sensitivity FDA-authorized SARS-CoV-2 RT-PCR test, run by trained laboratory personnel on a paired swab from the same subject. The population included symptomatic subjects and asymptomatic subjects with a recent exposure. The authorization also carried conditions requiring further evidence after authorization: a near-cutoff study showing that lay users can correctly identify low-positive samples, and a clinical evaluation supporting the serial-screening claim.

During the COVID-19 emergency, home molecular SARS-CoV-2 tests reached the market only through Emergency Use Authorization under section 564 of the Federal Food, Drug, and Cosmetic Act. FDA's expectations were set out in the Policy for Coronavirus Disease-2019 Tests and its EUA templates. The Molecular and Antigen Home Use Test Template called for a prospective clinical evaluation in the intended use population. Fresh samples were to be collected, tested and interpreted by lay users, compared against paired samples run on an authorized high-sensitivity RT-PCR comparator, at a minimum of two U.S. sites, at home or in a simulated home environment. The template recommended an observed usability study first, and a validated smartphone application where results are read or reported through an app.

The same template recommended NPA of at least 98% and, for OTC use in symptomatic individuals, PPA of at least 80%. For screening asymptomatic people, it recommended serial testing twice over three days, with at least 24 and no more than 48 hours between tests. A supplemental template allowed a serial-screening claim to be authorized on limited asymptomatic data, provided the developer committed to a post-authorization clinical study in asymptomatic individuals. Authorizations of this period often carried such conditions: data still owed to FDA after the test reached the market.

Home molecular tests add complexity that antigen tests do not have: a timed amplification step on a reusable heater, more handling steps and, for app-guided tests, software that interprets or records the result. Each step is a place where a lay user can go wrong. FDA's home-use design guidance treats these user-environment-device interactions as the central risk to manage.

• Having untrained subjects or parents run every step, including the app-guided reading, made the clinical PPA and NPA a measure of the whole system in real use, as FDA's home-use template recommends.
• Pairing a self-collected comparator swab with high-sensitivity RT-PCR run by trained laboratory staff gave a reference that was independent of the home workflow.
• Enrolling recently exposed asymptomatic subjects alongside symptomatic ones began to support the screening use. The post-authorization condition then addressed the full serial-screening claim.
• A blinded near-cutoff panel at about 2x LoD isolated the step most likely to fail in lay use, reading a weak positive line, without depending on how many low-viral-load positives happened to enroll.
• Tracing apparent false positives to their cause (result misinterpretation on an earlier app version) produced a specific, fixable design change, not an unexplained loss of specificity.