Rapid COVID tests are quite accurate at detecting infectiousness, correctly identifying roughly 80% of real infections when used with serial testing, but they are far less sensitive than laboratory PCR tests and can miss early infections entirely. Accuracy depends heavily on timing, symptom status, and how often you repeat the test. A single negative result is not proof you are uninfected.
The core problem is not the technology's design; it is the biology of when viral proteins appear and how much antigen a swab can capture. Understanding the mechanism explains why a test that is 90% accurate in one context behaves very differently in another. This guide breaks down the science behind rapid antigen tests, including the numbers, the edge cases, and the practical rules that actually matter.
How antigen tests detect the virus
Most rapid COVID tests are lateral flow immunoassays. A swab collects material from the nasal passages, and that sample is mixed with a buffer that breaks down mucus and stabilizes proteins. The mixture flows along a strip of nitrocellulose paper, driven by capillary action, and passes a line of antibodies that bind to the nucleocapsid protein of SARS-CoV-2 if it is present.
The detection chemistry is a sandwich format. Gold nanoparticles coated with one antibody attach to viral antigen in the sample. As the fluid moves, a second antibody fixed to the test line captures the complex, producing a visible colored line. A control line confirms the liquid flowed correctly, which is why an absent control line means the test failed rather than being negative.
The critical insight is that this assay measures antigen, not genetic material. It is a protein detection method, so it only works when enough viral protein is present in the sampled fluid. PCR amplifies nucleic acid through thermal cycling and can detect tiny quantities, which is why PCR remains the reference standard for infection status.
Two practical consequences follow from the mechanism. First, antigen tests are most accurate when viral shedding is high, which generally coincides with symptoms and infectiousness. Second, sample quality determines results as much as the reagent chemistry does; a shallow swab or a swab that misses the nasal turbinates can produce a false negative even with a good test.
Quantified accuracy in real populations
Head-to-head evaluations show a clear pattern. In peer-reviewed comparisons of rapid antigen tests against PCR in symptomatic people, the best-performing tests detect a large majority of infections, while sensitivity in asymptomatic people drops substantially. One widely cited comparison of several rapid tests found that sensitivity in symptomatic individuals ranged from about 60% to 80% for several dozen antigen tests, versus roughly 95% or better for laboratory PCR.
The date and symptom context matter. A positive result is highly informative because antigen tests are specific; when a test detects the nucleocapsid protein, it is unlikely to be a false positive from a random protein. A negative result is where the uncertainty concentrates. When a test detects an antigen, the probability that infection is real is high; when it does not, the probability depends on how many days have passed and how sick you are.
A useful way to frame this is to treat serial testing as the intended use case. Repeated testing every day or two during a symptomatic period catches many infections that a single test misses. A single negative test shortly after exposure has limited value; a negative test after several days of symptoms, combined with repeat testing, carries much more information.
| Test approach | What it detects | Typical accuracy context |
|---|---|---|
| Rapid antigen (single, symptomatic) | Viral nucleocapsid protein | High positive predictive value; sensitivity broadly around 0.7 to 0.9 against PCR in symptomatic adults |
| Rapid antigen (single, asymptomatic) | Viral nucleocapsid protein | Sensitivity substantially lower than in symptomatic people; best used serially |
| Rapid antigen (serial, twice over days) | Ranked likelihood of active infection | Sensitivity rises sharply with repetition across days |
| Laboratory PCR | Viral RNA (nucleic acid) | Reference standard; detects very low viral loads but can stay positive after infectiousness ends |
| At-home molecular test | Viral RNA (nucleic acid) | Near PCR sensitivity with slower turnaround |
| Antibody blood test | Host antibodies | Not a diagnostic test for current infection; indicates prior exposure |
The other useful number is specificity. Well-validated rapid tests report specificity in the high 90s to roughly 99% against PCR. In a low-prevalence population, even a tiny false-positive rate can produce a false result that looks positive, which is why confirmatory testing is sometimes recommended during periods of low community transmission.
The timing curve that explains most false results
Antigen tests fail mostly at the edges of the infection timeline. Infection requires the virus to replicate in the respiratory epithelium and produce enough nucleocapsid protein for the swab to carry measurable amounts. Early in infection, protein is present in small quantities, and the test may read negative. As the infection progresses and symptoms appear, shedding increases and the test becomes more reliable.
Here is the time course you should picture.
| Time point | What is happening in the body | What a rapid test typically shows |
|---|---|---|
| Day 0 after exposure | No viral replication or minimal localized infection | Negative is uninformative; too early |
| Days 1 to 3 | Early replication, rising antigen levels | False negatives are common |
| Days 3 to 7 | Peak viral shedding, onset of symptoms | Best detection window |
| Days 5 to 10 | High shedding, symptoms peaking | Strong detection; positive likely |
| Days 10 to 14 | Immunity rising, antigen declining | Test may turn negative before infectiousness fully ends |
| After day 14 | Viral protein largely cleared from the nose | Frequently negative even if symptoms linger |
The counterintuitive detail here is that a test can be strongly positive by day 5 and negative by day 12 even in the same person, without any treatment. The immune system clears antigen from the nasal mucosa faster than PCR clears viral RNA, so PCR can stay positive for weeks. That discrepancy produces confusion: a positive PCR two weeks after illness is not evidence you are still contagious.
Where rapid tests break down
Several edge cases change how a result should be interpreted.
False negatives are most likely very early in infection, when you are symptomatic but shedding below the detection limit, and when swab technique is poor. A negative test does not rule out infection on day one of symptoms, and it never rules out infection when you have a high-risk exposure without repeat testing.
False positives are rare but occur more often at low prevalence. A test is specific when it rarely lights up in uninfected people. In a period with little circulating virus, a positive result deserves a second test or a confirmatory molecular test before you make decisions about isolation or treatment.
Timing of the swab matters more than most people realize. The nasal turbinates, the ridges inside the lower nose, carry the highest antigen load. A swab that only brushes the nostrils misses them. Follow the manufacturer's instruction on how deep and how far to rotate the swab.
There is one professional caveat worth stating plainly. A negative test does not mean you are not contagious. Antigen tests correlate with infectiousness as a population-level tool, but individual shedding varies. If you have symptoms and a negative test, the most likely explanation is that you tested too early or sampled poorly, not that you are safe.
Common misconceptions worth correcting
A common belief is that a faint line means a weak result that should be ignored. A faint line is a real detection of antigen; the strength of the line does not reliably map to viral load or contagiousness, and any visible line should be treated as positive if it appears within the test's reading window. Reading a test after the stated window can produce lines that are not valid signals.
Another misconception is that a positive test proves you are infectious for the full isolation period. The test detects protein, and protein can remain detectable briefly after infectiousness peaks. What matters for infectiousness is active viral shedding, and the test is a rough proxy for that, not a direct measurement.
A third error is comparing a rapid test's accuracy directly with a PCR's accuracy. PCR is designed for sensitivity, so it detects tiny amounts of RNA. Rapid tests are designed for availability and speed, and their sensitivity is highest when antigen load is highest. They answer a different question: not "am I infected in any trace amount," but "is it likely I have an active, detectable infection right now."
A fourth is assuming antibody tests and antigen tests serve the same purpose. Antibody tests measure your immune response to a past exposure and are not used to diagnose current infection. They cannot tell you whether you are contagious today.
Practical steps and thresholds
Use rapid tests as a decision tool, not a single verdict. The following rules match the way public health agencies and manufacturers intend the tests to be used.
Test early and repeat. If you have symptoms, test on the first day and again in 24 to 48 hours. A single negative test on day one is weak evidence. Treat any visible test line within the reading window as positive. Ignore lines that appear after the window has expired. Treat a negative test as provisional. A negative result does not clear you if you have symptoms or a known exposure; it most likely means you tested too early. Use serial testing for exposure events. After a close contact, test on the day of exposure and again two to three days later. Confirm positive results with a molecular test when the test is being used for a decision that has consequences, such as returning to a high-risk setting. Prioritize the intended use case. Rapid tests are validated for self-collection in the nasal passages; do not substitute saliva or throat swabbing unless the manufacturer's instructions permit it.
These thresholds reconcile the two different numbers you will see in the literature. A single test may be around 70% to 90% sensitive in symptomatic people, but serial testing across several days brings the practical detection rate much higher. The published sensitivity figure applies to one test on one day; your real-world accuracy depends on repeat testing across days.
Limitations and what the science cannot tell you
Rapid antigen tests cannot reliably tell you whether you are negative. A negative result lowers but does not eliminate the probability of infection, especially in the first 48 hours of symptoms or after a high-risk exposure. Confirmatory testing, repeat testing, and symptom monitoring are still required for decisions that matter.
The tests also vary by manufacturer. Some have been evaluated extensively against clinical samples; others have less independent verification. The FDA's emergency use authorization process requires performance data, but the quality of evidence differs across products. Do not assume equal performance from equally packaged kits.
Finally, the accuracy data has boundaries. Most of the strong evidence comes from adults with symptoms and from comparing antigen tests with PCR, a reference standard that detects more infections than an antigen test will. The tests are also not validated to tell you when you are no longer contagious. What the evidence supports is a narrow, useful claim: rapid antigen tests are a good tool for identifying likely active infection when used repeatedly and in the right window, and they are least reliable when used once, at the wrong moment.