In January 2020, the entire world had exactly one laboratory capable of testing for SARS-CoV-2: the Chinese CDC in Beijing. By March, the US was processing fewer than 100 tests per day while South Korea was running 20,000. That gap wasn't a question of science. Both countries had the technology. It was a question of logistics, political will, and a testing infrastructure that had never been stress-tested at pandemic scale.
Understanding what each type of test actually does, and what it can't do, gives you a real advantage when the next outbreak hits. You'll know which result to trust, which to question, and when testing isn't even the right move.
How does PCR testing work?
PCR (polymerase chain reaction) detects a pathogen's genetic material by amplifying tiny fragments of DNA or RNA until they're measurable. It remains the gold standard for outbreak diagnosis because of its extraordinary sensitivity, capable of detecting as few as 10-100 viral copies in a sample. Results typically take 4-24 hours, though that window stretches dramatically when labs are overwhelmed.
A PCR test works in three steps. First, enzymes convert viral RNA into DNA (for RNA viruses like SARS-CoV-2 or influenza). Second, the machine heats and cools the sample repeatedly, doubling the target DNA sequence with each cycle. After 30-40 cycles, even a minuscule amount of viral genetic material becomes detectable. Third, fluorescent markers signal whether the target sequence was found.
PCR sensitivity typically exceeds 98% when the sample is collected correctly. Specificity approaches 100% because the test targets a genetic sequence unique to the pathogen. False positives are rare. False negatives happen, usually because the swab missed the virus or the patient was tested too early in infection before viral load built up.
Here's the catch: PCR can detect viral fragments for weeks after you're no longer infectious. During COVID-19, patients routinely tested positive 20-30 days after symptom onset, long after they'd stopped shedding live virus. A positive PCR tells you the genetic material is present. It doesn't tell you whether you're contagious right now.
What about rapid antigen tests?
Antigen tests detect proteins on the surface of the virus rather than its genetic material. They return results in 15-30 minutes and cost a fraction of PCR. During COVID-19, the BinaxNOW test retailed for about $24 for a two-pack compared to $100-150 for a PCR test at an out-of-network lab.
Speed and price come with a tradeoff. Antigen tests are less sensitive than PCR, typically catching 70-85% of infections. They perform worst during the first 1-2 days of infection and in people with low viral loads. A negative rapid test doesn't mean you're not infected. It means the test didn't find enough viral protein to trigger a positive signal.
Where antigen tests shine is in identifying people who are most contagious. High viral load produces a strong positive signal. If your antigen test shows a dark line within minutes, you're shedding a lot of virus. That correlation between test line intensity and contagiousness made rapid tests useful for real-time decisions during COVID-19: should you go to work today? Should you visit your elderly parent?
Serial testing improves accuracy. Two antigen tests taken 48 hours apart catch roughly 95% of infections, approaching PCR-level sensitivity at a lower cost and faster speed.
What do antibody (serology) tests tell you?
Antibody tests detect your immune system's response to a pathogen, not the pathogen itself. They identify IgM antibodies (which appear days to weeks after infection) and IgG antibodies (which develop later and persist for months or years). A positive antibody test means you were infected at some point. It says nothing about whether you're currently sick or contagious.
Serology's real value is population surveillance. During COVID-19, seroprevalence studies revealed that actual infections outnumbered confirmed cases by 5-10x in most countries. A February 2021 study in Manaus, Brazil found 76% seroprevalence, meaning three-quarters of the city had been infected despite official case counts capturing only a fraction. Those studies reshaped our understanding of COVID-19's true infection fatality rate, dropping it from the alarming early CFR estimates to the more accurate IFR.
You won't use an antibody test to find out if you're sick today. You'll use it to find out if you were sick last month, or whether a vaccine generated an immune response.
Why did COVID-19 testing collapse in 2020?
Quest Diagnostics and LabCorp, the two largest commercial lab networks in the US, process roughly 70% of the country's clinical lab tests. When COVID-19 testing demand surged in March 2020, both companies buckled. Quest reported average turnaround times of 5-7 days by mid-July 2020, with some results taking over two weeks. LabCorp wasn't much better.
A test result that arrives seven days later is almost useless for outbreak control. If you're infectious for 8-10 days and don't get your result until day 7, you've already spent most of your contagious period without knowing your status. Contact tracing based on week-old results is chasing ghosts.
Three factors created the bottleneck. Reagent shortages hit first, as global demand for RNA extraction chemicals outstripped supply from a handful of manufacturers. Equipment capacity was the second constraint; PCR machines can only run so many samples per batch. Staffing was third. Trained laboratory technicians can't be produced overnight.
South Korea avoided this collapse because it pre-positioned testing capacity after the 2015 MERS outbreak. The Korea Centers for Disease Control had pre-approved emergency-use protocols for novel coronavirus tests, enabling commercial labs to start producing kits within a week of the genome sequence being published. By early March 2020, South Korea had drive-through testing sites processing results in under 24 hours.
What do "sensitivity" and "specificity" actually mean?
Sensitivity measures how well a test catches true infections. A test with 95% sensitivity correctly identifies 95 out of 100 infected people and misses 5. Those 5 false negatives walk away thinking they're clean.
Specificity measures how well a test avoids false alarms. A test with 99% specificity correctly clears 99 out of 100 uninfected people and falsely flags 1. That 1 person gets a false positive.
Both numbers sound abstract until you consider prevalence. When only 1% of the population is infected, false positives start outnumbering true positives even with highly specific tests. Run a test with 99% specificity on 10,000 people where 100 are actually infected. You'll get about 95 true positives (assuming 95% sensitivity) and 99 false positives. Nearly half your positive results are wrong.
During low-prevalence periods, a positive rapid antigen test should be confirmed with PCR. During a surge when prevalence is high, a positive antigen test is almost certainly correct.
When should you test vs. when should you assume?
Testing makes sense when the result changes your behavior. If you have symptoms during an active outbreak and you're going to isolate either way, the test confirms the diagnosis but doesn't change your immediate actions. If you need a specific diagnosis to access treatment, as with Paxlovid for COVID-19 or Tecovirimat for Mpox, testing is necessary.
During peak outbreak waves, clinical diagnosis often replaces lab confirmation out of necessity. At the height of Omicron in January 2022, the UK stopped recommending confirmatory PCR after a positive rapid test. Hospitals in many countries diagnosed COVID-19 based on symptoms and chest imaging alone when testing capacity was exhausted.
Your decision framework is straightforward. Test when the result will change what you do: whether you isolate, which treatment you receive, or whether you travel. Skip the test when you're symptomatic during a known outbreak and plan to isolate regardless. Save limited testing capacity for situations where it matters most, particularly for people who are immunocompromised, elderly, or caring for someone vulnerable.
Point-of-care testing, performed at the bedside, pharmacy, or home rather than shipped to a central lab, is the direction the field is moving. Rapid molecular tests like Cepheid's GeneXpert and Abbott's ID NOW combine near-PCR sensitivity with 15-30 minute turnaround. As these platforms become cheaper and more widely deployed, the bottleneck that broke US testing in 2020 becomes less likely to recur. Less likely. Not impossible.